WO2025003604A1 - Device and method for the finite-element modelling of at least one part of the human or animal body - Google Patents

Abstract

The present invention relates to a computer-implemented method for the finite-element modelling of at least one part of the human or animal body, the method comprising the following steps: - defining a plurality of anatomical element control entities and a plurality of functional element control entities for an image representative of the at least one part of the human or animal body; - obtaining a mesh of the at least one part of the human body on the basis of the anatomical element control entities and the functional element control entities, wherein the anatomical element control entities and the functional element control entities are controlled during a geometric transformation of the mesh, the anatomical element control entities being positioned so as to define at least the geometric shape of the part and the functional element control entities being positioned so as to define mechanical connections associated with the part.

Description

Description

Title of the invention: device and method for finite element modeling of at least one part of the human or animal body

Technical Domain

[1] The present invention relates to the field of geometric and biomechanical models and more particularly to a method and a device for generating a geometric and biomechanical model in finite elements of at least one part of the human or animal body.

Previous technique

[2] The invention relates to the field of geometric and biomechanical models in finite elements of a structure consisting of all or part of the human or animal body. These models are virtual mock-ups which in particular make it possible to virtually exert localized forces or movements on the structure considered, and to deduce the mechanical response in terms of quantities of interest such as the amplitudes of mobility and mechanical constraints in specific areas. These structures can be damaged, and restored using orthopedic or surgical treatments, which treatments can give rise to mechanical complications, linked to overstresses in the implants or in the surrounding biological tissues.

[3] These models are often based on a single individual or an average individual considered representative. However, the great diversity of individuals requires that the model can be adapted to represent the form and function of the specific structures of different individuals. The construction of personalized models is important to understand which specificities have generated the mechanical over-constraints at the origin of the failure, and thus specify possible contraindications, or strategies adapted to certain specificities of the patient. However, creating the model for each of these situations can be particularly tedious, which is why these personalized models are very little developed today.

[4] We are particularly familiar with methods such as that disclosed by the document by Pei et al, published on January 19, 2023 and entitled “Biomechanical comparative analysis of conventional pedicle screws an cortical bone trajectory fixation in the lumbar spine: An in vitro and finite element study”.

[5] This method of building a biomechanical model is complex and tedious because it requires a lot of adjustments, and the example cited illustrates this complexity and its cost by the number of software that must be used for each step. This method is shown in Figure 1.

[6] This process includes

- a first step U1 of obtaining a series of scanner sections,

- a U2 segmentation step to obtain a 3D reconstruction,

- a U3 step of manipulation of the 3D reconstruction,

- a U4 step of meshing the implants and connecting elements,

- a U5 step of generating the finite element mesh,

- a U6 step of applying boundary conditions and integrating them into calculation software.

[7] The type of three-dimensional reconstruction obtained during step U2 may prove to be inconsistent on certain physical aspects, for example the interarticular facets may not be respected, which poses a problem in finite element models. These functional inconsistencies must therefore be corrected a posteriori, which in the example requires reworking the 3D reconstruction made before the finite element meshing process. In addition, mesh convergence processes are often required to validate finite element models, which requires generating, for the same structure, meshes with different mesh finesse, and this part (not carried out in the example) is often extremely tedious.

[8] The complexity of generating a finite element mesh from the external surface of the structures that compose it is one of the important obstacles that limit their development. Many methods have been proposed, which start from the envelope surface obtained.

[9] It is therefore necessary to propose new finite element modeling methods, making it possible to overcome all or part of these drawbacks.

Disclosure of the invention

[10] For this purpose, the present invention relates to a method of finite element modeling, implemented by a computer, of at least one part of the body. human or animal including the following steps:

- definition of a plurality of anatomical element control entities and a plurality of functional element control entities for an image representative of said at least one part of the human or animal body,

- obtaining a mesh of said at least one part of the human body from said anatomical element control entities and said functional element control entities, wherein said anatomical element control entities and said functional element control entities are controlled during a geometric deformation of said mesh, said anatomical element control entities being positioned on said representative image to define at least the geometric shape of said part and said functional element control entities being positioned on said representative image to define mechanical connections associated with said part.

[11] Thus, advantageously, the present invention can allow a saving of time and precision in the construction of the model. The present invention can be implemented to determine a representation of the human body taking into account the mechanical and morphological constraints. The present method can be used to allow obtaining a precise and rapid representation of the human body, to be used for example to prepare the manufacture of implants and to decide on their implantation. Taking into account the mechanical constraints, by the positioning of the control entities of functional elements can allow to better adjust the positioning of the implants for example. The present method can even be considered as a process which can be inserted into the manufacture of implants or prostheses.

[12] Indeed, taking into account functional element control entities to define mechanical connections associated with the part considered can accelerate the obtaining of a mesh compared to known methods. In addition, when the position of the mechanical connections is controlled, it does not move during the generation of meshes of different finesse for the convergence study and thus allows a significant saving of time and precision.

[13] According to at least one embodiment, the image representing said at least one part of the human or animal body represents

- an image obtained by medical imaging of said part or

- - a synthesized image of said part or - a parametric model generated from values of descriptors representative of the average of images representative of a plurality of patients.

[14] According to at least one embodiment, said plurality of anatomical element control entities and/or said plurality of functional element control entities are organized hierarchically, and respectively comprise:

- a first level of anatomical element control entities and/or a first level of functional element control entities:

- at least one second level of anatomical element control entities and/or at least one second level of functional element control entities, a higher level comprising respectively the functional element control entities of said lower level and additional functional element control entities,

- said mesh being obtained from a selection of said control entities of anatomical and functional elements.

[15] Thus, control entities can be chosen based on a desired mesh type. It may be relevant to be able to easily select them by correlating a desired mesh fineness at a hierarchical level.

[16] According to at least one embodiment, said selection selects at least one of said levels according to a mesh fineness to be obtained and/or an extent of the desired modeling in said part.

[17] According to at least one embodiment, said at least one part of the human body comprises at least one region of interest and said selection comprises a selection of a greater number of levels for said at least one region of interest than for said at least one part outside said region of interest, in order to obtain a mesh of which said fineness is finer for said at least one region of interest.

[18] It may be relevant to select more control points in regions representing the part(s) of the human body requiring in-depth detailed examination. For example, when inserting an implant, it may be relevant to obtain an accurate model in the region where the implant is inserted. More particularly, for example, when inserting a screw, it may be useful to obtain an accurate mesh at the screw insertion point. [19] According to at least one embodiment, the method comprises

- labeling, from said representative image of at least part of said plurality of anatomical and functional element control entities to obtain at least one location descriptor of said anatomical and functional element control entities,

- said mesh of said at least one part of the human body being obtained from said spatial coordinates.

[20] According to at least one embodiment, said mesh is obtained from a transformation of a reference mesh.

[21] According to at least one embodiment, said transformation is a deformation.

[22] According to at least one embodiment, said at least one location descriptor of at least a portion of said anatomical element control entities or said functional element control entities is obtained by a statistical estimation of a positioning of said anatomical element control entities or said functional element control entities.

[23] According to at least one embodiment, the method comprises a back projection of said mesh into said representative image.

[24] According to at least one embodiment, the method comprises

- an adjustment of at least part of said estimated location descriptors when their position in said mesh projected onto said representative image and their real position exhibit a difference greater than a determined value.

[25] According to at least one embodiment, said part comprises at least one osteoarticular or musculoskeletal segment consisting of one or more bone structures and associated connecting elements which can be instrumented by orthopedic implants.

[26] According to at least one embodiment, the method comprises

- the estimation, for at least one region of said part, of the bone mineral density from said representative image and

- the modification, in said obtained mesh, of mechanical properties associated with at least certain control entities of anatomical elements or said control entities of functional elements of said at least one region as a function of the estimation of bone mineral density. [27] According to at least one embodiment, the method further comprises obtaining a stiffness matrix.

[28] According to at least one embodiment, the method further comprises obtaining clinical indices from said obtained mesh.

[29] The generation of personalized finite element meshes with the proposed methodology can also greatly facilitate the automated calculation of many geometric and functional indices useful in particular for the clinician: these will be called clinical indices hereinafter for the sake of conciseness.

[30] The features presented in isolation in the present application in connection with certain embodiments of the method of the present application may be combined with each other according to other embodiments of the present method.

[31] According to another aspect, the invention also relates to a computer program comprising instructions for executing the steps of the method according to the invention, according to any one of its embodiments, when said program is executed by a computer.

[32] According to another aspect, the invention also relates to a computer-readable recording medium on which is recorded a computer program comprising instructions for executing the steps of the method according to the invention, according to any of its embodiments.

[33] According to another aspect, the invention also relates to a device for finite element modeling of at least one part of the human or animal body comprising one or more processors configured together or separately for the execution of the steps of the method according to the invention, according to any one of its embodiments. Thus, the present invention also relates to a device for finite element modeling of at least one part of the human or animal body comprising one or more processors configured together or separately for:

- defining a plurality of anatomical element control entities and a plurality of functional element control entities for an image representative of said at least one part of the human or animal body,

- obtaining a mesh of said at least one part of the human body from said anatomical element control entities and said functional element control entities, wherein said anatomical element control entities and said functional element control entities are controlled during a geometric transformation of said mesh, said anatomical element control entities being positioned on said image to define at least the geometric shape of said part and said functional element control entities being positioned on said image to define mechanical connections associated with said part.

[34] According to another aspect, the invention also relates to the use of a model obtained by a finite element modeling method according to the present disclosure to evaluate mechanical constraints brought about by modifications of the musculoskeletal structure of said human body.

[35] According to another aspect, the invention also relates to the use of a model obtained by a finite element modeling method according to the present disclosure to obtain clinical indices relating to said human body.

According to another aspect, the invention also relates to the use of a model obtained by a finite element modeling method according to the present disclosure for planning an orthopedic or surgical treatment to increase the efficiency and prevent failures related to implant placement.

Brief description of the drawings

[36] [Fig. 1] Figure 1 represents a method according to the prior art,

[37] [Fig. 2] Figure 2 shows a method according to one embodiment of the present invention,

[38] [Fig. 3a] Figure 3a shows an example of the positioning of the functional and anatomical control entities of level 1,

[39] [Fig. 3b] Figure 3b shows an example of the positioning of the functional and anatomical control entities of level 2,

[40] [Fig. 3c] Figure 3c shows an example of positioning of functional and anatomical control entities of level higher than 2,

[41] [Fig. 4a] Figure 4a represents an example of meshing according to a first meshing fineness of a lumbar segment according to an embodiment of the present invention, [42] [Fig. 4b] Figure 4b shows an example of meshing according to a second mesh fineness of a lumbar segment according to an embodiment of the present invention,

[43] [Fig. 5] Figure 5 shows an example of a parameterized representation,

[44] [Fig. 6] Figure 6 represents a mesh from the parameterized representation of Figure 5,

[45] [Fig. 7] Figure 7 shows a device according to one embodiment of the present invention.

Description of the embodiments

[46] The present disclosure concerns obtaining a geometric and biomechanical model. The models are widely used in the medical world and more particularly in surgeries involving implants and prostheses. Indeed, thanks to these models, it becomes possible to evaluate the mechanical constraints brought about by these modifications to the musculoskeletal structure. One of the aims being to avoid, a few months after the implant is placed, breakages or pain for the patient. However, current techniques for generating this biomechanical model remain ineffective because very often, their complexity of implementation does not allow individual adaptation to each patient. The present invention proposes an improvement in the generation of these biomechanical models, by allowing personalization by individual and therefore more realistic modeling, consideration of functional constraints from the start as well as speed of execution.

[47] The present invention is described mainly by taking the human body as an example but can of course be applied to the animal body.

[48] The method described in Figure 2 represents a finite element modeling method, implemented by a computer.

[49] This method may also be a finite element modeling method, implemented by a computer, to determine clinical indices relating to the human body modeled in finite elements at least in part.

[50] This method may also be a finite element modeling method, implemented by a computer, to plan an orthopedic treatment or surgical to increase efficiency and prevent failures related to implant placement.

[51] This method may also be a finite element modeling method, implemented by computer, to evaluate mechanical constraints brought about by modifications of the musculoskeletal structure of said human body or at least one part.

[52] This method can also be a finite element modeling method, implemented by computer, to determine a surgical strategy adapted to said human body.

[53] This method may also be a finite element modeling method, implemented by computer, to determine a mechanical response of the human body part to mechanical constraints. This may make it possible to deduce a mechanical response in terms of, for example, amplitude of mobility.

[54] For this purpose, the method comprises a first step E1 of obtaining a representation of one or more parts of the human body. It is envisaged that this is the entire human body. By part of the human body, one or more components is generally understood. Preferably, the part comprises at least one osteoarticular or musculoskeletal segment consisting of one or more bone structures and associated connecting elements. In the case where the model is obtained for use in the context of surgery, the part can also be instrumented by orthopedic implants or prostheses.

[55] By representation is meant an image, or a plurality of images, obtained by medical imaging of the part of the human body or a modeling of a part of the human body. Different medical imaging systems can be used and in particular scanners, magnetic radiation imaging (MRI), plane or bi-plane X-ray systems, ultrasound imaging systems. Of course, it may be envisaged to use a plurality of images with a plurality of systems, taken from different points of view or angles, of the part of the human body.

[56] By modeling, we can understand a synthesized image of said part, the synthesis being able for example to be representative of a population model. By modeling, we can also speak of a parametric model, obtained not from the image of a patient, but from values of descriptors representative of the average of images from different patients. It is also possible to generate data artificially for special cases such as unusual dimensions: for some vertebrae for example, it is possible to vary the thickness of the pedicles (or any other dimensions) to carry out sensitivity studies aimed at studying the effect of a specific factor, or to highlight specificities that some patients may have, in search of worst case scenarios.

[57] According to certain embodiments, to obtain the representation, it is possible to estimate the data of a patient from his external envelope, by statistical modeling such as that proposed in the article by A Nerot et Al: Estimation of spinal joint centers from external back profile and anatomical landmarks » published in the « journal of Biomechanics » on March 21, 2018, pages 96 to 101.

[58] It is important to note that the present method can be much more accurate for a given patient when the representation(s) used in subsequent steps are derived from a real image of the body part and not an image obtained by modeling.

[59] An optional calibration step (not shown) can be considered when the representation is an image from a medical imaging device.

[60] The calibration may be of the geometric type according to certain embodiments. It may be necessary to move from a two-dimensional image to a three-dimensional image. This calibration is necessary for example in images from a scanner or magnetic resonance imaging or biplanar radiographs. A registration may also be necessary when the images come from different imaging modalities, taken in different positions, the mesh may for example be defined from the image obtained from the scanner (in a lying position) and readjusted on another image representative of the same part of the body, for example an RX image taken in a standing or sitting position. The calibration parameters may also be estimated from the potential position of the imaging device and the subject or from a calibration object whose dimensions are known.

[61] The calibration may also be a grayscale calibration according to some embodiments. This is particularly advantageous for customize the mechanical properties. Indeed, bone mineral density is correlated with the mechanical properties of the bone. The variation in bone mineral density is observed by variations in the gray level of the bone. Thus, calibration methods adding a so-called calibration phantom object to the image can be used to determine bone mineral density. More approximate methods, using self-calibration assuming known bone mineral densities of very dense regions and of low-density regions can also be used.

[62] In the present disclosure, the notion of control entities of anatomical or functional elements is associated with a control point in general but can also be associated with a geometric primitive, such as ellipses, circles or others.

[63] In a step E2, control entities are defined. This step may or may not be broken down, for clarity, into two sub-steps E21 and E22 but these two sub-steps may be carried out at the same time, simultaneously or one before the other, in any order.

[64] In step E21, anatomical element control entities, also called anatomical control entities for simplification, are defined, determined or obtained. The anatomical element control entities may also be called morphological or geometric element control entities. These anatomical element control entities make it possible to define the geometric shape of at least one part of the human body. They are not necessarily representative of mechanical connections of this part or of this part associated with another part or with an external element, but may be useful for characterizing clinical indices used by clinicians.

[65] These anatomical element control entities consist of points or sets of points allowing the shape of the part of the human body to be reconstructed in a sufficiently precise manner. The number of points or sets of points determines the desired precision.

[66] Anatomical control entities can be defined, determined, or obtained, manually or automatically, in several ways. Thus, for example, some can be determined manually by an operator. Intelligent software based on learning, such as artificial intelligence, can also be used to define, or determine or obtain the control entities on the representation. When it is desirable to determine several tens to several hundreds of entities, or even more, the control points can advantageously be used to generate a first personalized mesh from which can be automatically adjusted, iteratively or directly, points of higher levels whose adjustment is required, and this thanks to a method described later, based on the personalized mesh from one or more salient areas of the part. An advantage of the method is that only the control points of the required level are to be adjusted, there is no need to adjust thousands of points which are of no interest from a clinical point of view or from a finite element modeling point of view.

[67] In the case where the human body part is a spinal segment, the anatomical control entities may be points or sets of points representative of the centers of the endplates or their contour, the centers of the pedicles or specific points of the posterior arch.

[68] In a step E22, functional element control entities are defined, determined or obtained. It may be noted that for reasons of clarity, the determination of the functional element control entities is indicated in a step E22. However, the determination of the functional element control entities may be done simultaneously with the determination of the anatomical element control entities, for example in step E2. The functional element control entities are defined to define mechanical connections associated with the part of the human body for which they are defined. By associating we can mean the intra mechanical connections to this part or inter mechanical connections, connecting this part to at least one other part of the human body or connecting this part to an external part, such as an implant, which will be inserted into this part or into this part and another part, connecting them for example.

[69] By means of such anatomical and functional element control entities, the present disclosure can enable the obtaining of more relevant and personalized models for an individual and the obtaining of which is much faster than the methods of the prior art. In the case where the part of the human body considered is a spinal segment, the functional element control entities can be positioned at the insertion points of certain fibers of the intervertebral disc, or ligaments, points or attachment zones of the ligaments above spinous, interspinous, intertransverse for example. The functional element control entities can also be related to the implants to be inserted and more precisely can be placed at the entry point of the implant attachment screws, or at the location of cages.

[70] The definition of functional control entities can allow the creation of the finite element mesh to take into account the insertion points of the ligaments, the fibers of the intervertebral disc or even the entry points of the pedicle screws, for example.

[71] Generally speaking, the determination and placement of these functional element control entities advantageously make it possible to place in a controlled manner points or sets of points other than points with purely geometric purposes, to identify points of the part of the human body that must not move, either during the deformation of the mesh (obtained during subsequent steps), or when a finer mesh is constructed. When the part of the human body comprises multi-bone structures, these functional entities are particularly relevant for modeling inter-articular functional constraints.

[72] The control entities of functional elements can be defined, determined, or obtained, manually or automatically, in several ways. Thus, for example, some can be determined manually by an operator. Intelligent software based on learning, such as artificial intelligence, can also be used to define, or determine or obtain the control entities on the representation. This is conceivable for a few entities but when it is desirable to determine several tens to several thousands of entities, the points can be determined automatically, without human intervention, by image processing software which automatically determines the contours or salient areas of the part.

[73] In the case where the part of the human body is a spinal segment, the functional element control entities can be attachment points or zones of the supraspinous, interspinous, intertransverteral ligaments. The functional control entities can also be linked to the placement of implants and represent an entry point for the implant attachment screw. The definition of the functional control entities allows the consideration for the creation of the finite element mesh, of the insertion points of the ligaments, of the fibers of the disc intervertebral or even pedicle screw entry points. The methods of the prior art relying only on anatomical control entities, induce insufficiencies in mesh control. Indeed, the generic models created in the prior art were built based on the adjustment constraints of the geometry of an isolated bone of complex shape, without considering the fact that it is a polyarticulated system, and that in addition to the points which allow the geometry to be deformed, the functional points associated with the connecting elements should also be taken into account. The present disclosure, in contrast to the known methods, relies on an intelligent mesh, comprising control entities of anatomical and functional elements of osteoarticular segments, consisting of one or more bone structures and associated connecting elements, intact, damaged and/or instrumented by orthopedic implants. The present invention thus makes it possible not to consider bones for example in isolation but to take into consideration the connections linking this bone to the adjacent parts to which it is linked, by taking into account for example interarticular or ligament connections. The functional control entities can be representative of these connections.

[74] Functional element control entities and anatomical element control entities are entities whose position is controlled and does not move during the convergence study of a mesh, which involves the construction of finite element models with different mesh finesse. The position can be controlled by adjusting a location descriptor (e.g. of the spatial coordinate type) so that the control entities are positioned to correspond to the morphology of the part of the human body of the individual concerned. By morphology, we can mean the shape but also the position of the mechanical connections between the different structures making up the part, such as, for example, ligaments, bones, muscles, etc.

[75] Other points or entities may also be located manually or automatically without constituting control entities, in order to adjust the mesh as determined in the following steps. The positioning of these points is not the subject of the present disclosure.

[76] According to certain embodiments, the entities obtained during step E2 can be organized hierarchically and can respectively comprise: - a first level of entities for controlling anatomical elements and/or a first functional element control entity level:

- at least one second level of anatomical element control entities and/or at least one second level of functional element control entities, a higher level comprising respectively the functional element control entities of said lower level and additional functional element control entities.

[77] The first level entities are considered as the minimal entities that allow to generate the finite element mesh obtained during the following steps. Entities of level N, N greater than 1, ... can possibly be determined in order to refine the finite element mesh described below. For example, the mesh of a vertebral segment can, depending on the desired precision requirement, be represented by a few tens, hundreds or thousands of points, and the control entities can vary from a few entities to a few hundred entities.

[78] Level 1 anatomical and/or functional element control entities may be those that are essential regardless of the desired mesh fineness to describe the essential dimensions and to describe the location of the essential anatomical or functional element control entities. Level 2 control entities may include, in addition to the level 1 control entities, entities representing additional details useful in certain applications or when customization of certain regions is important. Other sub-levels, 3, 4, or even more, may also be defined.

[79] In some embodiments, examples of entities controlling level 2 or higher anatomical elements within a spinal segment include:

- points belonging to vertebral walls, transverse processes or laminae (other than functional points on these surfaces), whose need for precision may, in a number of applications, appear marginal. It will therefore not be necessary to spend time reconstructing them precisely, which speeds up the mesh generation time.

[80] According to some embodiments, there may be cited, as examples as level 1 functional element control entities,

- the points or areas of attachment of the supraspinous and interspinous ligaments, the centers facets, etc.

- the screw entry point for a screw-longitudinal element type implant.

[81] According to certain embodiments, it is possible to cite, as an example as entities for controlling anatomical elements of level N greater than 1, within the framework of a spinal segment:

- specific points delimiting the connection zone between the pedicles and the vertebrae,

- lateral points of intersection of the plates with the frontal plane of the vertebra.

[82] According to certain embodiments, there may be cited, as examples as entities for controlling functional elements of level 2 or 3, within the framework of a spinal segment:

- main attachment points of the intervertebral disc fibers on the endplates,

- control points describing for example one or more posterior or anterior points of the plateau at the intersection with the sagittal plane of the vertebra, which can control the positioning of the posterior and anterior longitudinal ligaments

- a point associated with the length or orientation of a screw,

- the main dimensions and orientation parameters of the plates, pedicles or facets from a parameterized representation,

- points allowing to control the thickness of cortex in a given zone.

[83] Figures 3a to 3c show examples of positioning of these anatomical and functional element control entities.

[84] In Figures 3a to 3c, the anatomical element control entities are represented by crosses, the level 1 functional element control entities by circles and the level 2 functional element control entities by squares.

[85] Figure 3a represents a positioning of the functional element control entities of level 1. These primary functional element control entities are positioned on the geometric centers of each vertebral endplate of the segment considered.

[86] Figure 3b shows a positioning of the functional element control entities of level 1 or primary, represented by the circles and level 2 or secondary, represented by the squares. These functional element control entities Secondary functional joints are positioned at the insertion points of the ligaments and at the centers of each articular facet.

[87] Figure 3c shows a positioning of the functional element control entities of level 1, represented by the circles and of level 2 or secondary, represented by the squares and of control entities. These secondary functional element control entities are positioned on the contours of vertebral plates to control the insertion points of a part of the fibers of the intervertebral disc. Anatomical element control entities are also positioned on the contours of the plates to control the shape of the plates, their orientation and the heights of the discs.

[88] During a step E3, one or more desired or to-be-obtained mesh models can be defined, or one or more variants of mesh models can be defined from a set of meshes with the same control points and different mesh finenesses. One or more finenesses, also called resolutions, of mesh to be obtained can thus be specified. The mesh fineness can be defined between 1 and N, N greater than 1, the larger N, the finer the mesh. The desired mesh fineness can be linked to the application of the model, or to the part of the human body considered, certain parts or regions of interest in a part requiring a mesh of higher fineness. The fineness of the mesh can also be defined in relation to the surgery envisaged on said part or region of interest following the obtaining of the model. It can also be linked to the performance of a convergence study to compare meshes of different finesse.

[89] Among the types of mesh that can be selected, we can notably mention hexahedral type meshes.

[90] According to some embodiments, the mesh can be created from the control entities and a parameterized representation. A representation in high-order isoparametric elements can also be envisaged, making it possible to obtain a representation of a complex shape with a quality hexahedral mesh. According to some embodiments, this parameterized representation can be associated with that described in the document by Lavaste et al entitled "Three-dimensional geometrical and mechanical modeling of the lumbar spine". published on October 25, 1992. [91] Figures 4a and 4b illustrate for example a spine model with two different mesh finesses, which can be adjusted using geometric transformations to position the control points in subject-specific X, Y, Z coordinates.

[92] Figure 6 illustrates a model of blocks predefined by reference cubes, which can then be assembled using geometric transformation functions to position the control points in subject-specific X, Y, Z coordinates.

[93] In a step E4, a mesh, called a reference mesh, is selected from the meshes preselected in step E3. This reference mesh advantageously allows, in the subsequent steps E8 to E10, the adjustment of the location of the functional and anatomical element control entities estimated in step E7. Thus, the meshes selected in step E3 other than the reference mesh are not used, or are not necessary or used, for adjusting the location of the functional and anatomical element control entities estimated in step E7. The selection criterion is the mesh with the smallest number of nodes which favors the adjustment of the location of the estimated control entities: this selection depends on the available representations: for example, when the representation requires producing DRRs (Digitally reconstructed radiographs), it may be interesting to choose a mesh with sufficient resolution to produce realistic DRRs.

[94] According to certain embodiments, it may be noted that steps E3 and E4 may be replaced by a single step in which not several models are selected from which a reference mesh is selected but immediately the reference mesh.

[95] During a step E5, functional element control entities and anatomical element control entities can be selected. This selection can advantageously make it possible not to use all the previously defined anatomical and functional element control entities. This selection can therefore advantageously make it possible to reduce the number of control entities used for the subsequent steps of the method and thus to reduce the calculation times required during the subsequent steps. This selection can be made according to a mesh fineness to be obtained in the subsequent steps and/or an extent of the desired modeling in the part considered. By extent, it is possible meaning that a selection of a greater number of control points can be made for an area of the human body part, such as a region of interest. For example, a greater number of points can be selected for a region of interest of the human body part under consideration than for areas outside that part.

[96] As mentioned with respect to step E2, the functional element control entities may be organized hierarchically.

[97] The selection may consist of selecting the control entities of anatomical and/or functional elements by hierarchical level. The selection may consist of selecting a greater number of levels for a region of interest than for areas outside the region of interest. This can result in a finer mesh for the region of interest.

[98] Selection may also involve selecting more control entities of anatomical and/or functional elements for a region of interest, even when the control entities are not hierarchically organized.

[99] According to certain embodiments, a region of interest is considered or defined that can more precisely represent an area of the part of the human body on which the model will be used for a medical application, for example an area that a practitioner wishes to examine, consider, for example to insert an implant. For this region of interest, higher level control entities can be selected compared to the rest of the part of the human body. Thus, subsequently, a finer mesh can be obtained for this region of interest. In addition, having more control entities selected, there will be more precision, while keeping acceptable calculation times since high level control entities are not selected for the entire part.

[100] In some embodiments a region of interest may be the disc adjacent to spinal instrumentation.

[101] Steps E1 to E5 may correspond to a first phase in which the control entities may be defined, these being able to condition a family (or library) of meshes, as well as a mesh model.

[102] The following steps, E6 to E10, may constitute a second phase in which the spatial location of the control entities is estimated for a given individual. Steps E6 to E10 represent a first embodiment of the phase 2. According to this embodiment, the model is personalized for a given individual, by refining the positioning of the control entities for this individual, during steps E6 to E10.

[103] During a step E6, selected anatomical and functional element control entities may be labeled. The labeling may make it possible to obtain, from the representation, the placement of the control entities obtained during the previous steps on the representation of the part of the human body. To this end, the labeling may make it possible, from the representation, to obtain at least one location descriptor of the control entities on the representation and according to certain embodiments, spatial coordinates of at least a portion of the anatomical and functional element control entities. In some embodiments, the labeling may also allow for associating an anatomical identification with the control entities of anatomical elements (e.g. the end of the spinous process of each vertebra) and functional elements (e.g. the insertion points of the interspinous ligaments, linking the spinous process of each vertebra to the spinous process of the overlying vertebra on the one hand and of the underlying vertebra on the other hand.

[104] The term coordinates is used instead of location descriptor when the control entities are represented by points and in the remainder of the description, control entities represented by points are taken as examples and coordinates are therefore associated with them. But in general, the term coordinates can be replaced, throughout the description, by location descriptor.

[105] In some embodiments, the labeling information is stored as a file.

[106] According to certain embodiments, this labeling can be carried out manually, from medical images, from automatic annotations from the images based on artificial intelligence techniques, or from semi-automatic techniques.

[107] According to certain embodiments, this labeling can be constituted from anatomical and/or statistical considerations, possibly using a parameterized description of the object, adjusted to the overall dimensions of the individual. [108] According to some embodiments, in some representations, it is possible that certain functional or anatomical element control entities cannot be placed, because they are not visible on the representation. In this case, their coordinates can be estimated by different statistical methods, for example regression methods, or methods called "posterior shape modeling", or by taking advantage of transverse inferences, by describing the relationships between dimensions of the same vertebra, or longitudinal inferences, by describing the relationships between the dimensions of the vertebrae of the same segment for example.

[109] According to some embodiments, the control entities of anatomical or functional elements are not represented by points and are therefore not associated with coordinates but can take other geometric shapes, and for example the shape of an ellipse. In such an embodiment, the descriptors can be the spatial coordinates of the center, the values of the major axis and the minor axis of the ellipse, and the spatial orientation parameters of the plane of the ellipse.

[110] At the end of step E6, the location of the control points on the representation is available, in the form of a location descriptor which can take different forms, for example coordinates but also include any other information suitable for allowing the placement of the control entity on the representation.

[111] During a step E7, a labeling of the additional control entities is estimated with respect to the control entities selected during step E5. This estimation can advantageously be done automatically using intelligent software based on different approaches, such as those based on statistics and/or learning, and artificial intelligence. These approaches allow a rapid estimation, particularly advantageous when it is desired to use a large number of control entities that would be tedious to label manually. This estimation can consist of an automatic labeling of these additional control entities. These additional control entities can be control entities of one or more hierarchical levels higher than the control entities labeled during step E6. This therefore advantageously makes it possible to benefit in the following steps from a plurality of control entities for obtaining the mesh and therefore of a more precise mesh, while avoiding having to manually label a very large quantity of control points.

[112] In a step E8, a mesh of the human body part is obtained. The mesh is obtained from the anatomical element control entities and the functional element control entities, and the reference mesh model, in which the anatomical element control entities and the functional element control entities are controlled during a geometric transformation of the mesh, the anatomical element control entities being positioned to define at least the geometric shape of the part and the functional element control entities being positioned to define mechanical connections associated with the part.

[113] According to certain embodiments, the obtained mesh can also be obtained from the desired mesh model when the latter is defined, for example, during step E3 and when it is not previously defined by default.

[114] As mentioned previously, this mesh can be obtained from a selection of control entities and in particular from a selection by hierarchical levels of control entities.

[115] According to a first embodiment, this mesh can be obtained from and in particular by transformation of an existing mesh.

[116] According to certain embodiments, this transformation is a deformation of an existing mesh. There are today a plurality of prior art techniques describing a mesh deformation. Preferably, the mesh is of the hexahedral type and its deformation can for example be of the “Kriging” type. in certain embodiments the transformation can comprise an operation called mesh quality control and “regularization”, as described in the document “Jacobian-based repair method for finite element meshes after registration” by Bucki M., Lobos C., Payan Y., published in the journal “Engineering with Computers 2010”.

[117] A mesh consists of a set of structured points comprising connections between the points. A particularity of the present invention is to obtain a mesh in which functional element control entities are immediately taken into account for obtaining the mesh, as controlled elements of the mesh, as are the control entities of elements anatomical. Prior art meshing methods used only anatomical element control entities. These functional element control entities are controlled anchor points that do not move during mesh deformation or during subsequent mesh refinement steps for mesh convergence studies and thus do not need to be repositioned when it is necessary to study mesh convergence by subjecting the model to several calculations with different mesh finenesses. Thus, a time saving as well as a gain in accuracy are obtained by defining the functional element control entities upstream of obtaining the mesh.

[118] Figures 4a and 4b illustrate examples of meshes of a lumbar vertebral segment obtained from in particular the control entities of anatomical and functional elements of Figure 3c. Thus, the mesh of Figure 4a has a less fine mesh resolution (or finesse) than the mesh of Figure 4a. The meshes of the two figures are however based on the same control entities of anatomical and functional elements but the mesh obtained in Figure 4b is finer. These two meshes may have been defined in phase E3, and correspond to a generic geometry. The reference mesh for the adjustment of the estimated control entities may be, depending on the need, that shown in Figure 4a or in 4b. Once the location of the geometric and functional control entities has been adjusted, it is possible to automatically produce a resolution 1 mesh (RI type) or resolution type 2 (R2 type), or any other type that relies on the same control entities. It is also possible to select only a subset of the control entities to produce a less fine mesh in all or part of the vertebral segment.

[119] Thus, the present disclosure allows the obtaining of a finite element model upstream of a three-dimensional reconstruction from the representation. This finite element model immediately constitutes the generic object to be deformed for the three-dimensional reconstruction.

[120] At the end of step E8, the mesh of the part of the human body can be back-projected, from the 3D space associated with the image or series of images (for example a pair of biplanar X-rays or a set of CT slices), onto the representation of this part during step E9. This back-projection has the following purpose: objective, if necessary, to verify the location of the estimated control entities, to possibly allow their adjustment described in E10.

[121] It should be noted that for the purposes of this backprojection the finest mesh can be considered because it does not impact the finite element calculation cost at this stage, and facilitates automated adjustments. Since the model is hierarchical, the finest model contains models of a lower level, which makes the transition from one finesse to another almost automatic. This is interesting because in some applications we may be interested in having a global representation in some parts, and much more detailed in another (for example the disk adjacent to spinal instrumentation).

[122] This may involve customization based on the individual's specific features. Thus, it may be considered to obtain a topography of the finite element mesh of the structure of interest (the part of the human body observed) that is identical regardless of the individual considered, when for example the initial representation was based on a modeling of this structure of interest. The additional step E10, optional, may allow a more precise approach to adjust the location of the control entities estimated in E7 to adapt to the individual concerned more precisely.

Thus, during step E10, at least part of the location descriptors, or spatial coordinates estimated during step E7, can be adjusted when their position in said mesh projected onto said representation and their actual position have a difference greater than a determined value. The adjustment can be manual, and carried out for example by an operator. Image processing methods (contour detection, etc.) can be used to automate this replacement of points that are not correctly placed. When the initial placement of some of these points (whose location descriptor is obtained by estimation) is too far from the position where they should be, then several iterations of steps E8 to E10 may be necessary, by adjusting the nearby points to loop back with a generation of a closer mesh (because it includes the points already adjusted) and a backprojection of this mesh for a new adjustment of the points not yet adjusted. In these iterations, the reference mesh is generally the same.

[124] The images can indeed contain artifacts: for example when there is metal in scanner sections, the neighboring areas are noisy, the facets interlocking joints may look glued. Building a finite element mesh with reliable and clearly visible information on the image, and supplementing with an estimate of the missing information, while respecting functional constraints (non-interpenetration of facets for example) can be very effective.

[125] The iterations end when the previously estimated control entities are correctly placed, or in other words labeled, on the part of the human body considered, that is to say when they correspond to representative points of the anatomy for the control entities of anatomical elements or to mechanical connections when they correspond to control entities of functional elements. The iterations can also be terminated from other stopping criteria, for example after a limited number of iterations, considering that beyond that the gain in precision is marginal. Thus, it is possible to obtain a placement of the control entities from a single reference mesh. In other words, steps E8 to E10 are steps that can allow an adjustment of the location (or a determination of the coordinates) of the control entities that can then allow one or more meshes of a different type (or finesse for example) to be obtained, from these correctly positioned control entities.

[126] The iterations possibly carried out during steps E8 to E10 are iterations which do not require a new adjustment of the positioning of the control entities made during step E10 and this or these iterative adjustments are therefore very fast, almost instantaneous, compared to the adjustments which were necessary in the methods of the prior art which did not take into account the functional control entities during a progressive adjustment of the initial mesh. These iterative steps do not require mesh regularization because the transformations are small. It is possible, if necessary, to have quality control and regularization at step Eli.

[127] As mentioned previously, the steps E6 to E10 described previously are an embodiment of the phase of adjusting the control entities adapted to the morphology of an individual. According to another embodiment, it is also possible not to determine a precise positioning of the control entities per individual but to take into consideration an average individual or even to take a most unfavorable case (also called "worst case"). case scenario » in English), or a model representative of a category of individuals with a specific morphotype. In such an embodiment, the positioning of the control entities is not carried out by successive iteration as in steps E6 to E10 described but the positioning of the control entities is determined from a file comprising the location descriptors (or coordinates) of the control entities of functional and anatomical elements.

[128] During an Eli step, once the anatomical element control entities and the functional element control entities are correctly placed, a mesh is obtained from the meshes determined during step E3, and from the correctly placed control entities. Thus, a mesh is obtained for each type of mesh selected during step E3. If during step E3 a single mesh model is defined, then during the Eli step, a single mesh is obtained. This Eli step may consist of generating the mesh from the correctly positioned control entities. It may include a quality control and mesh regularization operation as specified previously.

[129] It is therefore observed that taking into account functional constraints, by positioning functional control entities from the creation of the mesh, makes it possible to obtain an adapted mesh very quickly. Unlike the methods of the prior art, the present method therefore uses an “intelligent” object to perform the three-dimensional reconstruction of the part of the human body analyzed. The mesh is particularly intelligent for two main reasons, namely:

- each point of the mesh can be localized on the topological plane, which can facilitate the process of deformation of the generic object to adapt to the images by only considering the useful points to finally arrive at the desired finite element mesh, this structuring, with the localization of each point, can facilitate the calculation of clinical indices

- the finite element mesh is constructed in a physically consistent manner (for example, the interarticular facets are congruent), with some existing geometric transformation methods being able to take this congruence into account.

[130] Thus, by obtaining the finite element model constituting the generic object to be deformed for three-dimensional reconstruction, a quality finite element mesh can be obtained. One of the main advantages is to make the This agile and efficient process, in particular for meshing structures composed of several components, by controlling not only the geometry but also the functional aspects. We thus have, from the geometric reconstruction phase, the advantages of a finite element mesh, which for a bone part of the human body, pushes regionalization very far, by integrating the functional surfaces useful on the biomechanical level, and which has interesting qualities in terms of mesh regularity. In addition, when the representation takes into account several objects, that is to say when the part of the human body includes several objects linked together, such as bones with articular or ligamentous segments, the essential constraints in finite elements, such as for example, the non-interpenetration of articular surfaces, are immediately satisfied, avoiding an additional cost when the corrections are to be made at the end of the three-dimensional reconstruction. In addition, it is possible to integrate into this generic model a generic distribution of bone mineral density (BMD). Due to the existing correlation equations between DMO value and mechanical properties (Young's modulus and maximum admissible stress), it is then possible to attribute a generic distribution of mechanical properties to the model.

[131] When said representation is a grayscale representation and calibrated in bone mineral density, thanks to a calibration phantom, an equation is available to transform the grayscale information (Hounsfield units) into BMD (bone mineral density) information. Due to the relationships between BMD and mechanical properties, it may then be directly possible to adjust the mechanical properties of each bone element.

[132] When the mesh is fitted to an individual using planar or multiplanar radiographs, the benefit of such a distribution is to have more realistic digitally reconstructed radiographs (DRRs), which then facilitates the adjustment of the position of certain control points.

[133] The following steps may constitute a plurality of applications of the preceding steps.

[134] For this purpose, according to a first application, at the end of the meshing process, a calculation step, E12, can make it possible to implement the meshes obtained and to apply to them boundary conditions (in forces and in displacement) linked to the application of the model. It should be noted that this calculation step, applied to each Selected meshes allows to select the mesh that is sufficient for the application to which the method will be applied. This selection can further take into consideration requirements of standardization documents, such as ASME V&V40 (Assessing Credibility of Computational Modeling through Verification and Validation: Application to Medical Devices, The American Society of Mechanical Engineers), in particular the requirement to perform a mesh convergence study by performing the simulations with meshes of different finesse.

[135] At the end of step E12, it is still possible to define a new mesh by adjusting additional points that are not control entities as defined previously, namely control entities defined in the hierarchical representation and which characterize a family of meshes, but which are points in variable number in meshes of the same family, and which one may wish to adjust in a specific application. This can be done following a backprojection of the mesh onto the image, the control entities already positioned correctly and adjusting additional points there. This new selection can relate to one or more areas of interest of the part of the human body or its entirety. The control entities having already been positioned, there is advantageously no longer any need to adjust the control entities but only the additional points. This can allow to obtain a more exact mesh in an area of interest (for example the points which define a pedicle when the finite element study is focused on the study of the screws which are inserted there).

[136] According to certain embodiments, the finite element mesh as obtained at the end of step E12 can be used to assist in the design of implants. For this purpose, during the step of defining the functional element control entities, step E2, it is possible to position primary and possibly secondary functional element control entities and more at the location of insertion of the screws for example. The construction of the finite element mesh can therefore take these points into account for the construction of the initial mesh and does not require them to be positioned on an initial mesh obtained solely from anatomical control entities. This considerably improves the speed of the method since the iterations of steps E8 to E10 are minor adjustments compared to the adjustment that would have been necessary if the functional control entities had been positioned after obtaining the mesh based solely on the anatomical control entities. [137] According to certain embodiments, the finite element mesh as obtained at the end of step E12 can be used for clinical use in the service of implant designers for the design of their products, or in the service of a surgeon for the planning of a surgery. Indeed, with this technique, the model can quite naturally be adapted to the patient to be treated, because the initial representation can be made from one or more medical images of the patient and therefore, the adjustments made during steps E8 to E10 are minor. It is also possible to start from a predetermined model, probably generating more adjustments during steps E8 to E10, depending on the morphology of the patient and the difference between his morphology and the typical morphology of the predetermined model used.

[138] The method can be used to plan an orthopedic or surgical treatment to increase efficiency and prevent failures related to implant placement. The positioning of some of the functional control points at the location of screws, for example, can be used to see, after obtaining the model, following step E12, whether the selected location is suitable for the patient in view of their particularities, including their morphology and the different pathologies. Adaptation can be understood to mean:

- the exposure of said model to constraints,

- observation of the behavior of the model under constraints,

- repositioning of control entities and generation of a new mesh (see step E12) to determine the best position for the implant.

[139] It is therefore easy to understand that the implant designer or surgeon finds it in their interest to be able to benefit from a rapid finite element meshing method in order to be able to test different configurations.

[140] The easy creation of the finite element mesh of a large number of patients, integrating the definition of the functional element control entities associated with the placement of implants, thus makes it possible to construct the finite element models of these patients and to deduce therefrom the mechanical constraints linked to the placement of implants. These calculations, on a series of patients, thus make it possible to specify whether for certain patients, mechanical overstresses, causes of failures, are likely to appear, and thus to verify the safety of the implant and possibly specify contraindications. This approach by large-scale digital simulation, greatly facilitated by the present invention, can make it possible to limit the use of real clinical trials, which is of considerable interest given the high costs of a clinical trial campaign and the human cost represented by possible failures that could have been identified in silico instead of on the patient. Similarly, in the case of monitoring patients following implant placement or surgery, it is important to identify cases in which a mechanical complication occurs. The construction of a personalized model such as that proposed by the present invention can make it possible to understand which specificities have generated the mechanical over-stresses at the origin of the failure and thus specify possible contraindications or strategies adapted to certain specificities of the patient. Thus, the present method can advantageously allow the planning of an orthopedic or surgical treatment to identify the strategy best suited to a given patient to increase effectiveness and prevent such failures.

[141] The construction of personalized models is important to understand which specificities have generated the mechanical over-constraints at the origin of the failure, and thus specify the possible contraindications, or strategies adapted to certain specificities of the patient.

[142] The method may also be a computer-implemented finite element modeling method for evaluating mechanical stresses introduced by modifications to the musculoskeletal structure of said human body or at least one part thereof.

[143] Another example of application is the application to the definition of clinical indices. Thus, in addition to allowing integration into a finite element calculation, the mesh obtained can be directly used for the calculation of clinical indices, such as the orientation of the vertebral plates, the dimensions of the pedicles, or the bone mineral density in a region of interest. These indices are useful for diagnostic purposes or for evaluating the effect of treatments. It should be noted that the 3D reconstructions obtained by the classic segmentation methods from series of CT or MRI slices give access to point clouds that do not correspond to anatomical or functional "labels", which does not allow the desired indices to be obtained directly, which highlights the advantage of the proposed method.

[144] For this purpose, following step Eli, a step E13 can be carried out allowing the definition of clinical or functional indices translating the geometric relationships or functional between the entities. Thus, the method makes it possible to determine clinical indices relating to the human body modeled in finite elements at least in part.

[145] Clinical indices can be defined for each of the custom meshes in step Eli, when several custom models have been defined in step E3.

[146] A third application example can be used to determine an associated stiffness matrix.

[147] The method as shown in Figure 2 also comprises steps E14 to E16, which may constitute a second aspect of the invention. It may be noted that steps E14 to E16 are numbered 14 to 16 but are not sequential or performed following step E13.

[148] From step E3, it is possible to obtain a definition of the stiffness matrix associated with each selected mesh. The constitution of the stiffness matrix is the first step performed by the finite element calculation software, to constitute the matrix system allowing the analysis of the behavior of a structure under the effect of specific boundary conditions (imposed displacements and loads). By solving the system of linear equations associated with the stiffness matrix, the displacements, deformations and stresses and deformations in the structure can be determined. Since the topology of the mesh is the same for the generic mesh and for the personalized mesh, it is possible, as soon as a mesh is defined during step E3, to initialize the stiffness matrix associated with this mesh.

[149] This stiffness matrix includes the generic mechanical properties and the generic coordinates of the points.

[150] This stiffness matrix is adjusted during a step E15, in the last phase of backprojection and adjustment of the ECs, following steps E8 to E10 (or iteratively during loops E8-E10). This adjustment can concern the nodes of the mesh, and also the mechanical properties from the gray levels on the image.

[151] During a step E15, we can solve the system.

[152] The advantage of such a method of obtaining the stiffness matrix is twofold: - On the one hand, a time saving because there is no new matrix construction pass in the calculation software,

- On the other hand, this faster step allows you to then move directly to the resolution of the system of equations and is therefore a gateway to a fast and efficient dedicated solver, which represents an important issue given the cost of general finite element calculation software, and the growing need to reduce simulation times or even achieve real-time simulations.

[153] It should also be noted that, in order to save time, the adjustment of the stiffness matrix could only be carried out on the areas that impact the desired results. Also, since the boundary conditions (in displacements and forces) are the same for all individuals, it is possible to preprocess this matrix from phase E14 to take these boundary conditions into account, which makes it possible to speed up the adjustments and the resolution. Finally, such an approach lends itself well to new resolution methods concerning large-dimensional sparse matrices (which is generally the case for these matrices) by fast iterative methods.

[154] Figure 5 shows a mesh model that can be used in step E3, in the form of a parameterized representation used as a model of predefined blocks by reference cubes, which can then be assembled using geometric transformation functions to position the control points in the X, Y, Z coordinates specific to the subject's human body part and thus to the subject-specific characteristics.

[155] In this representation, each cube is represented in a different size allowing to find a little the shape of the vertebra, but in practice each cube is a reference element of side 2. In a reference cube for an isoparametric formulation, the coordinates of the ends are limited to (-1, 1) in the three directions of space, which amounts to having a cube of side 2. In these elementary cubes defined in a reference frame (r,s,t) the mesh can be very easily controlled (ensuring the coherence of the links between cubes, see figure 6), and the geometric and functional control points can be defined. The isoparametric transformation functions allow to transform each cube according to the coordinates (X, Y, Z) of specific points, to arrive at a parameterized mesh of this surface. In this figure the following names are used: CV: vertebral body P: pedicles PAC: central articular processes PAS: superior articular processes PAI: inferior articular processes PT: transverse processes L: laminae PE: spinous processes.

[156] Figure 6 shows an example of meshing applied to the parametric representation of Figure 5.

[157] Figure 6 illustrates a mesh of a plurality of elements represented in Figure 5 and in particular the vertebral body, CV, the superior articular process PAS, the central articular process PAC, the transverse process PT.

[158] A is an example of a vertebral body mesh. The vertebral body is represented as a reference cube with center (0,0,0) in the (r,s,t) coordinate system, r, s or t taking the values from -1 to 1. In this cube, the mesh control can be defined very easily, so A and A' represent two different mesh finesse. The geometric control points PI, P2, P3, P4 allow the connection between A (or A ⁷ ) and the pedicle B (or B'). With different mesh resolutions, PI, P2, P3, P4 remain the same. Note that the pedicle is also represented in the form of a reference cube with center (0, 0, 0) in the (r,s,t) frame, r, s or t taking the values from -1 to 1. These reference elements are as defined in classical works, in finite elements and when formulations in isoparametric elements are mentioned, for their ease in defining both the displacement fields and the geometric transformations, with functions of a higher or lower order. This is illustrated for example in the document

"Numerical methods in finite element analysis", by K. - J. Bathe and E. L. Wilson, published in "Prentice - Hall Inc., Englewood Cliffs, N.J" in 1976.

[159] In an original way, it is used here to obtain a hexahedral mesh of controlled fineness in each block representing the subdivisions of each of the objects constituting the structure to be meshed, by taking advantage of the geometric transformation functions allowing to pass from the local reference points of each reference cube to the global reference point X, Y, X common to all the cubes. In a way original, these reference cubes are enriched with the location (in local coordinates r, s, t) of the geometric and functional control points, which makes it possible to obtain an estimate of their location X, Y, Z as soon as the appropriate geometric transformation is applied to each of the reference cubes.

[160] C gives an example of an articular process represented by a reference cube. D represents the reference cube of the transverse process. Points L1 and L2 are functional control points allowing the insertion of the transverse ligaments. Points Ql, Q2, Q3, Q4 are geometric control points allowing the connection between C and D.

[161] E represents the reference cube in which 9 functional control points Fl, F2, F3, F4, F5, F6, F7, F8, F9 allow the contact connection with the facet of the upper vertebra.

[162] In the present example the reference elements considered are cubes, of course other types of reference elements (surface, beams, as described for example in Bathe), could be used if necessary, considering a higher or lower order for the geometric transformation functions depending on the need.

[163] Figure 7 schematically represents a preferred embodiment of a device 1 capable of implementing a method according to an embodiment of the present disclosure, typically the method described in Figure 1.

[164] The device 1 comprises one or more processors 11 implementing a method according to the present disclosure, a read-only memory 12 (of the “ROM” type), a rewritable non-volatile memory 13 (of the “EEPROM” or “NAND Flash” type for example), a rewritable volatile memory 14 (of the “RAM” type) a communication interface 15 with for example a medical imaging device of the scanner or magnetic resonance type. The read-only memory 12 constitutes a recording medium in accordance with an exemplary embodiment of the invention, readable by the processor or processors 11 and on which is recorded a computer program PI in accordance with an exemplary embodiment of the invention. Alternatively, the computer program PI is stored in the rewritable non-volatile memory 13.

[165] The computer program PI may enable the device 1 to implement at least part of the method according to the present disclosure. [166] This computer program PI can thus define functional and software modules, configured to implement the steps of an information transmission method in accordance with an exemplary embodiment of the invention, or at least part of these steps. These functional modules rely on or control the hardware elements 11, 12, 13, 14, 15 of the device

1 cited above.

[167] The device 1 may be, in a non-limiting manner, a computer, a server, a tablet, a “smartphone” type telephone.

[168] The present invention can be applied to the field of transport (in particular automotive with virtual crash tests) and to the field of sport. Indeed, one of the fields of application is modeling for the analysis of injury mechanisms during impacts (automobiles, pedestrians, sports, ...).

Claims

Claims [Claim 1] A method of finite element modeling, implemented by a computer, of at least one part of the human or animal body comprising the following steps: - definition of a plurality of anatomical element control entities and a plurality of functional element control entities for an image representative of said at least one part of the human or animal body, - obtaining a mesh of said at least one part of the human body from said anatomical element control entities and said functional element control entities, wherein said anatomical element control entities and said functional element control entities are controlled during a geometric transformation of said mesh, said anatomical element control entities being positioned on said image to define at least the geometric shape of said part and said functional element control entities being positioned on said image to define mechanical connections associated with said part. [Claim 2] A method according to claim 1 such that said representative image of said at least one part of the human or animal body represents - an image obtained by medical imaging of said part, or - a synthesized image of said part or - a parametric model generated from values of descriptors representative of the average of images representative of a plurality of patients. [Claim 3] Method according to one of claims 1 or 2 in which said plurality of anatomical element control entities and/or said plurality of functional element control entities are organized hierarchically, and respectively comprise: - a first level of anatomical element control entities and/or a first level of functional element control entities: - at least one second level of anatomical element control entities and/or at least one second level of functional element control entities, a higher level comprising respectively the functional element control entities of said lower level and additional functional element control entities, - said mesh being obtained from a selection of said control entities of anatomical and functional elements. [Claim 4] Method according to claim 3 in which said selection selects at least one of said levels according to a mesh fineness to be obtained and/or an extent of the desired modeling in said part. [Claim 5] A method according to claim 4 wherein said at least one part of the human body comprises at least one region of interest and said selection comprises a selection of a greater number of levels for said at least one region of interest than for said at least one part outside said region of interest, in order to obtain a mesh whose fineness is finer for said at least one region of interest. [Claim 6] Method according to one of the preceding claims comprising - labeling, from said representative image of at least part of said plurality of anatomical and functional element control entities to obtain at least one location descriptor of said anatomical and functional element control entities, - said mesh of said at least one part of the human body being obtained from said at least one location descriptor of said control entities of anatomical and functional elements. [Claim 7] Method according to one of the preceding claims in which said mesh is obtained from a transformation of a reference mesh. [Claim 8] A method according to claim 7 wherein said transformation is a deformation. [Claim 9] The method of claim 6 wherein said at least one location descriptor of at least a portion of said anatomical element control entities or said functional element control entities is obtained by a statistical estimation of a positioning of said functional element control entities. control of anatomical elements or said functional element control entities. [Claim 10] Method according to one of the preceding claims comprising a back projection of said mesh into said representative image. [Claim 11] A method according to claims 9 and 10 comprising - an adjustment of at least part of said estimated location descriptors when their position in said mesh projected onto said representative image and their real position exhibit a difference greater than a determined value. [Claim 12] Method according to one of the preceding claims such that said part comprises at least one osteoarticular or musculoskeletal segment consisting of one or more bone structures and associated connecting elements which can be instrumented by orthopedic implants. [Claim 13] Method according to one of the preceding claims as it comprises - the estimation, for at least one region of said part, of the bone mineral density from said representative image and - the modification, in said obtained mesh, of mechanical properties associated with at least certain control entities of anatomical elements or said control entities of functional elements of said at least one region as a function of the estimation of bone mineral density. [Claim 14] Method according to one of the preceding claims such that it further comprises obtaining a stiffness matrix. [Claim 15] Method according to one of the preceding claims such that it further comprises obtaining clinical indices from said obtained mesh. [Claim 16] Computer program comprising instructions for executing the steps of the method according to one of claims 1 to 15 when said program is executed by a computer. [Claim 17] A computer-readable recording medium on which is recorded a computer program comprising instructions for carrying out the steps of the method according to one of claims 1 to 15. [Claim 18] Device for finite element modeling of at least one part of the human or animal body comprising one or more processors configured together or separately for: - defining a plurality of anatomical element control entities and a plurality of functional element control entities for an image representative of said at least one part of the human or animal body, - obtaining a mesh of said at least one part of the human body from said anatomical element control entities and said functional element control entities, wherein said anatomical element control entities and said functional element control entities are controlled during a geometric transformation of said mesh, said anatomical element control entities being positioned on said image to define at least the geometric shape of said part and said functional element control entities being positioned on said image to define mechanical connections associated with said part. [Claim 19] Use of a model obtained by a finite element modeling method according to one of claims 1 to 15 for evaluating mechanical constraints brought about by modifications of the musculoskeletal structure of said human body. [Claim 20] Use of a model obtained by a finite element modeling method according to one of claims 1 to 15 to obtain clinical indices relating to said human body. [Claim 21] Use of a model obtained by a finite element modeling method according to one of claims 1 to 15 for planning an orthopedic or surgical treatment to increase efficiency and prevent failures linked to implant placement.