RU2848435C2 - Method for morphological processing of radar images in the microwave range in the field of medicine using various hypotheses regarding the environment through which microwave signals pass - Google Patents

Abstract

FIELD: medicine.

SUBSTANCE: the invention relates to a method and device for processing medical images of human tissues in the patient's body, and in particular in the mammary gland, using a medical imaging device (1) containing a group of microwave probes, wherein said group contains K > 1 probes separated by gaps from each other, the group of probes has P > 1 different configurations defining transmitting probes and receiving probes for one or more positions around the area under consideration, wherein the transmitting probes are designed to transmit microwave range signals to irradiate the body region, and the receiving probes are designed to receive microwave range signals after they have been scattered and reflected in the region of interest, in an additional variant, the probes may be designed to transmit and receive simultaneously.

EFFECT: improved quality of radar images for analysing human tissues or organs.

7 cl, 12 dwg

Description

Field of technology to which the invention relates

The present invention relates to the field of medical imaging using ultra-high frequency (UHF) electromagnetic waves and, more specifically, to medical imaging for the analysis of human tissues or organs permeable to electromagnetic waves. The present invention finds particular application in breast imaging and the detection of breast pathologies.

State of the art

Microwave medical imaging technologies enable imaging of human organs that are transparent to electromagnetic waves and represent promising technologies in the field of breast medical imaging and the detection of pathologies such as breast cancer.

Microwave imaging utilizes transmitter probes capable of irradiating the entire organ to be imaged, or a portion thereof, with electromagnetic waves. The transmitted waves pass through the region to be imaged and are then received by receiving probes. These probes can optionally be configured to transmit and receive simultaneously. The received waves pass through the region to be imaged and are reflected by obstacles encountered in areas where there is a change in permittivity (e.g., a cancerous lesion located in healthy tissue). All transmission coefficients measured between the transmitter and receiver probes constitute a multistatic data set. These multistatic data sets serve as input to radar image processing modules and enable the generation of a two-dimensional (2D) or three-dimensional (3D) radar image of the entire organ or a portion thereof.

To obtain images that best represent the area to be imaged, it is necessary to know a priori the characteristics of the dielectric medium along the various paths taken by electromagnetic waves between the emitting probes, each point in the area to be imaged, and the receiving probes.

However, such a priori knowledge of the dielectric properties of the organs to be visualized is not available and requires assumptions regarding the characteristics of the medium being imaged, resulting in images that may be of poor quality.

Disclosure of the essence of the invention

The present invention enables the improvement of radar image quality for the analysis of human tissues or organs. Specifically, the present invention enables the acquisition of radar images of one or more regions of interest corresponding to physical objects that may represent diseased areas.

To this end, the present invention provides, according to a first aspect, a method for processing medical images of human tissue in a region of a patient's body, and in particular the mammary gland, by means of a medical imaging device comprising a group of probes emitting/receiving electromagnetic waves in the microwave range, wherein the group consists of K> 1 probes separated by gaps from each other, the group of probes has P> 1 different configurations defining emitting probes and receiving probes for one or more positions around said examined region, wherein the emitting probes are configured to emit microwave signals in order to irradiate the region of the body, and the receiving modules are configured to receive microwave signals after their scattering and reflection on the irradiated region, the probes can, in an additional embodiment, be configured to simultaneously emit and receive signals, the method comprises the following steps, implemented in a processor module of the medical imaging device:

- collect signals generated in P > 1 antenna group configurations;

- the method for each configuration contains the following steps:

○ process signals collected for N > 1 sets of Ai > 1 values, 1 ≤ I ≤ N, characteristics of the human tissue parameters through which the signals under consideration passed, to obtain elementary microwave radar images;

○ selecting in each of the N sets, in accordance with at least one image focusing criterion, an elementary microwave radar image, wherein each selected elementary image corresponds to one of the values of the parameter (pcfib) from the specified set; wherein one elementary image from each set is selected for the configuration;

- the method for each set contains the following steps:

○ reconstruct, based on the selected elementary radar images, each of the radar image configurations of the patient's body region under consideration to reconstruct one three-dimensional (3D) radar image for each set;

○ Morphological processing is performed on each reconstructed radar image to obtain a morphological image in which one or more regions of interest, if present, that may represent the affected areas are identified

- evaluate the consistency of the presence of each of the regions of interest in the various morphological images obtained in order to morphologically confirm the existence of the region of interest.

The present invention is preferably supplemented by the following characteristics, taken individually or in any of the technically possible combinations:

- morphological processing consists of determining the density of one or more pixel regions of an image; the region of interest is identified if the density associated with it exceeds a threshold value,

- take into account at least N=2 sets, preferably at least N=3 sets of Ai > 1 values, 1 ≤ i ≤ N, of the characteristic parameter (pcfib) of the medium through which the signals pass,

- these sets completely or partially overlap one another in terms of ranges of change and/or in terms of values,

- the assessment of the persistence of presence consists of determining the percentage of presence of the region of interest in the morphological images under consideration; the region of interest is considered confirmed if this percentage of presence exceeds a certain threshold value,

- a group of probes contains K > 1 probes located around the area whose image is to be obtained, said group is configured to move between vertical positions around the area whose image is to be obtained, each configuration represents an angular sector of the group of probes, so that each sector contains M > 1 probes, where M < K, each configuration is shifted in the angular direction by at least one probe from any other configuration; in each sector, each probe emits in turn to generate signals collected for this configuration, an elementary radar image is determined for each angular sector in a set of values of the characteristic parameter.

The present invention provides, according to a second aspect, a computer program product comprising program code instructions for performing the steps of the method according to the first aspect of the present invention, where the method is performed by at least one processor.

The present invention provides, according to a third aspect, a medical imaging device comprising a processing module adapted to perform the method according to the first aspect of the present invention.

The combination of several configurations and the use of several characteristic values of the medium through which the radiation passes as assumptions makes it possible to work with an a priori unknown inhomogeneity of the dielectric properties of the region whose image needs to be obtained.

Furthermore, the consistent presence of a region of interest identified morphologically for at least B > 1 sets of values out of N sets (where B/N ≤ 1) implies confirmation of the region of interest, in other words, a high probability that the region of interest corresponds to a physical object/lesion actually present in the imaged region and not an artifact.

In an additional variant, the density criterion used is a shape descriptor that allows for successful morphological identification of regions of interest.

Description of drawings

Other features, objects and advantages of the present invention will emerge from the following description, which is purely illustrative and not limiting, and which should be read in conjunction with the accompanying drawings, where:

Fig. 1 is a simplified illustration of a medical imaging system using microwave radiation according to one embodiment of the present invention;

Fig. 2 illustrates the steps of the method for morphological processing of microwave radar images according to the present invention;

Fig. 3 illustrates a microwave radar image obtained using the method according to the present invention;

Fig. 4a, Fig. 4b, Fig. 4c, Fig. 4d, Fig. 4e illustrate morphological images of the patient's mammary gland obtained using the method of morphological processing of radar images according to the present invention.

Fig. 5a, Fig. 5b, Fig. 5c, Fig. 5d illustrate morphological images of the patient's mammary gland obtained using the method of morphological processing of radar images according to the present invention.

In all drawings, similar elements have the same reference designations.

Implementation of the invention

Fig. 1 illustrates a device 1 for medical imaging using microwave radiation, comprising an examination table 11 on which a patient 12 lies. In particular, the patient 12 lies on her stomach. The examination table 11 preferably comprises a circular opening 13, allowing the patient's mammary gland 14 to be immersed in a reservoir 15 filled with a biocompatible matching fluid, the dielectric properties of which are optimized for transmitting electromagnetic waves into the mammary gland.

A group of 16 probes 161 emitting/receiving microwave electromagnetic waves (hereinafter schematically shown in dashed form) is arranged around the reservoir 15 and allows, in the emission mode, to irradiate the observed medium and, in the reception mode, to receive signals reflected from the scene, the image of which is to be obtained. The probes 161 are preferably uniformly distributed around the reservoir and are preferably arranged in the form of a ring surrounding this reservoir, as shown in Fig. 1. Preferably, the probes are configured to emit signals in the frequency range of 0.5 - 6 GHz.

More generally, the imaging system operates in multistatic mode and allows the medium to be imaged to be illuminated using multiple probes in emission mode and multiple probes in reception mode, with the probes forming various configurations around the medium to be imaged. Alternatively, the probes can be configured to simultaneously emit and receive radiation.

For each multistatic data set, all or part of the environment to be imaged is sequentially irradiated using preselected probes operating in emission mode. The number of probes operating in emission mode is selected from a group based on the breast region to be imaged. The signal from each emitting probe is received by preselected probes operating in reception mode. The number of probes operating in reception mode is selected from a group based on the breast region to be imaged. Each multistatic data set is then considered to correspond to a series of signal emission/reception operations by probes in a given configuration.

By configuration, we mean the definition of a plurality of emitting probes and the definition of a plurality of receiving probes that allow the collection of a set of multistatic data regarding the entire mammary gland or a part thereof, these probes being positioned in a certain manner in the space around the mammary gland.

To switch from one configuration to another and to control the collection of various sets of multistatic data, the system comprises a module 17 for controlling a group of probes, connected to a module 18 for monitoring and processing data (e.g., a processor or a computing device). Such a module 18 for monitoring and processing data is configured to control a group of sensors, to perform data collection, to ensure the storage of the collected data, to perform radar image processing operations and to implement a method for morphological image processing, which will be described below. A storage module 19 makes it possible to store all the collected multistatic data and a certain amount of data that can be used in the image processing stages or created as a result of image processing. In addition, a display module 20 makes it possible to present and view the obtained images. The module 18 for monitoring and processing data, the storage module 19 and the display module 20 can be integrated directly into the visualization device or located physically separately from it. Image processing operations can be performed a posteriori (offline).

As will be clear, several sequential configurations of transmitter and receiver probes are defined for imaging the entire breast. These transmitter and receiver probe configurations cover different areas of the breast to be imaged and are selected to ultimately cover the entire breast to be imaged.

For each configuration, multistatic data sets containing the transmission coefficients between the emitting probes and receiving probes make it possible to obtain, after processing the radar images, an elementary radar image.

All of these elementary images allow for the reconstruction of a two-dimensional (2D) or three-dimensional (3D) radar image of the visualized region, in this case, the breast. For information on radar processing of multistatic transmission/reception signals, allowing for the reconstruction of two-dimensional (2D) or three-dimensional (3D) radar images, see, for example, the following publications:

- A.J. Devaney, Time reversal imaging of obscured targets from multistatic data, IEEE Trans. Antennas Propag. (2005). doi:10.1109/TAP.2005.846723;

- Marengo, E.A.; Gruber, F.K.; Simonetti, F. Time-reversal MUSIC imaging of extended targets. IEEE Trans. Image Process. 2007, 16, 1967-1984. doi:10.1109/TIP.2007.899193);

- Hossain, M.V.; Mohan, A.S. Cancer Detection in Highly Dense Breasts Using Coherently Focused Time-Reversal Microwave Imaging. IEEE Trans. Comput. Imaging 2017, 3, 928-939. doi:10.1109/TCI.2017.2737947);

- A. Fasoula, B.M. Moloney, L. Duchesne, J.D.G. Cano, B.L. Oliveira, J. Bernard, M.J. Kerin, Super-resolution radar imaging for breast cancer detection with microwaves: the integrated information selection criteria, in: 41st Annu. Int. Conf. IEEE Eng. Med. Biol. Soc., 2019.

The resulting two-dimensional (2D) or three-dimensional (3D) radar images are preferably used as part of a data processing method that will be described below.

As noted in the introduction, defining each elementary radar image theoretically requires a priori knowledge of the characteristics of the dielectric medium in which the mammary gland is located (in other words, the medium through which electromagnetic waves propagate) along the path between each transmitting probe and each receiving probe. However, this information is unavailable.

As will be described below, the present invention implements the pcfib parameter, which corresponds to the assumption of the average composition of the medium through which electromagnetic waves pass in the mammary gland (or in the imaged region as a whole), in terms of permittivity. This pcfib parameter corresponds to the percentage of the mixture of fibroglandular tissue and adipose tissue in the mammary gland. For example, pcfib = 30% corresponds to a medium containing 30% fibroglandular tissue and 70% adipose tissue. The dielectric properties of the mammary gland tissue are then determined as a weighted average (weighted using the pcfib parameter) of the dielectric properties of the fibroglandular tissue and adipose tissue. For example, information on the permittivity values of fibroglandular tissue and adipose tissue in the mammary gland can be found in the following publications:

- T. Sugitani, S.I. Kubota, S.I. Kuroki, K. Sogo, K. Arihiro, M. Okada, T. Kadoya, M. Hide, M. Oda, T. Kikkawa, Complex permittivities of breast tumor tissues obtained from cancer surgeries, Appl. Phys. Lett. (2014). doi:10.1063/1.4885087;

- M. Lazebnik, L. McCartney, D. Popovich, K.I. Watkins, M.J. Lindström, J. Harter, S. Suval, F. Magliozzo, J.H. Busquet, M. Okuniewski, S.K. M. Lazebnik, L. McCartney, D. Popovic, C.B. Watkins, M.J. Lindstrom, J. Harter, S. Sewall, A. Magliocco, J.H. Booske, M. Okoniewski, S.C. Hagness, A large-scale study of the ultrawideband microwave dielectric properties of normal breast tissue obtained from reduction surgeries, Phys. Med. Biol. (2007). doi:10.1088/0031-9155/52/10/001.

The method of morphological processing of microwave radar images is described below in connection with Fig. 2.

First, P > 1 probe group configurations are determined (step E0) in such a way as to be able to cover the entire breast to be imaged and then to reconstruct a three-dimensional (3D) radar image of this breast.

Then, for each configuration, a multistatic data set is collected containing the transmission coefficients measured between the transmitting and receiving probes (step E1). This results in several multistatic data sets (P > 1 multistatic data sets).

The signals collected for each configuration are then processed to produce elementary microwave radar images for each configuration (step E2).

In particular, to process these signals, several sets (N > 1 sets) of Ai values (Ai > 1 values, where 1 ≤ I ≤ N) of the parameter pcfib are taken into account. Then there are values of the pcfib parameter for each configuration. Thus, based on the signals of each set of multistatic data, we obtain elementary microwave radar images, each obtained for a single value of the pcfib parameter. The idea here is to obtain elementary images under different assumptions regarding the characteristics of the medium through which the electromagnetic waves pass.

Preferably, these sets of pcfib parameter values overlap each other, completely or partially, in terms of changes and/or values.

For example, you could have one set containing values of 10%, 20%, and another set containing values of 5%, 15%, and 25%. In this example, you have a set whose values vary between 10% and 20%, and another set whose values vary between 5% and 25%. Thus, these two sets have a total range of variation between 10% and 20%.

In another example, you could have a set containing the values 10%, 20%, and another set containing the values 20%, 25%, 30%. In this example, the specified sets have one common value of 20%.

Another example might be to have a set containing values of 10%, 20%, 25%, and another set containing values of 5%, 10%, 30%. In this example, the two sets have a common range of variation between 10% and 25% and a common value of 10%.

Here, at least two sets of pcfib parameter values are considered, one of which may have a wider range of pcfib parameter value variations than the other set. Here, the terms "wide" and "narrow" are relative, understood by comparing the ranges of variation. The idea here is to have overlap between the sets of values.

The choice of the ranges of variation of the pcfib parameter for different kits is made in accordance with the existing variability in terms of composition and density of mammary glands.

Preferably, wide ranges of change result in images containing a more complete representation of the region of interest, while narrow ranges of change potentially result in partial representations of the detected lesions.

For example, in the context of breast imaging, N=5 change sets may be selected:

- three sets with narrow ranges of change:

○ between 10% and 20%, the pcfib parameter takes, for example, the following values in this range: 10%, 15%, 20%

○ between 30% and 40%, the pcfib parameter takes, for example, the following values in this range: 30%, 35%, 40%

○ between 50% and 60%, the pcfib parameter takes, for example, the following values in this range: 50%, 55%, 60%

- two sets with wide ranges of changes:

○ between 20% and 50%, the pcfib parameter accepts, for example, the following values in this range: 20%, 25%, 30%, 35%, 40%, 45%, 50%

○ between 10% and 60%, the pcfib parameter takes, for example, the following values in this range: 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 55%, 60%.

For each configuration, one elementary microwave radar image is selected for each set (step E3), corresponding to one of the parameter values (pcfib) from this set; thus, for each configuration, one elementary image is selected for each set. This selected image may contain a region of interest corresponding to a physical object potentially representing the affected area. In the previous example, there are then five elementary images for each configuration (one elementary image for each set), which will be used for reconstruction.

Such a choice consists, in particular, of using at least one of the image focusing criteria (image focusing metric parameters) such as, for example, the criteria described in the following documents (a combination of several criteria may be used):

- S. Pertuz, D. Puig, M.A. Garcia, Analysis of focus measure operators for shape-from-focus, Pattern Recognit. (2013). Doi :10.1016/j.patcog.2012.11.011).

- O’loughlin, D.; Krewer, F.; Glavin, M.; Jones, E.; O’halloran, M. Focal quality metrics for the objective evaluation of confocal microwave images. Int. J. Microw. Wirel. Technol. 2017, 9, 1365–1372. Doi :10.1017/S1759078717000642).

Each selected image may contain one or more regions of interest, which are selected according to at least one image focusing criterion. Indeed, the selection is accomplished using one or more focusing criteria. For example, for a selected criterion, this procedure involves selecting the image that yields, for example, the minimum value of a metric parameter from the set of all others for that criterion.

Note that, from configuration to configuration, the choice of an elementary image for the same set under consideration can be carried out for different values of the pcfib parameter belonging to this set.

Based on the elementary images obtained for different configurations, a two-dimensional (2D) or three-dimensional (3D) radar image of the visualized region is reconstructed for each set (step E4). Thus, for each set of pcfib values, one radar image is reconstructed.

Each microwave radar image reconstructed in this manner is subjected to morphological processing to identify regions of interest, if any (step E5). The resulting image, called a microwave morphological image, contains none or one or more of the identified regions of interest. Such morphological processing consists, in particular, of identifying connected objects in the image using a thresholding method and of preserving as regions of interest those connected objects that correspond to a set of morphological characteristics, in particular, the volumetric size of the connected object, the density level of this connected object, the intensity level within the connected object, the level of contrast between the intensity within the connected object in question and the intensity within other connected objects potentially identified in the same image. At this stage, there are several morphological images of the breast, each morphological image being obtained for one set of pcfib values; each morphological image contains none or one or more of the identified regions of interest.

Preferably, morphological processing is based on a density criterion. Density is calculated as the ratio between the volume of the object and the volume of the object's convex hull. Generally, the greater the density of a region of interest, the more this region of interest will be "filled" (an area of interest without openings), the more it will have a well-defined and convex outline, and therefore the more likely it is to correspond to a mass. In the case of breast imaging, this density criterion may be intended to support the differentiation between a breast mass and a focal asymmetry (an "island" of normal breast tissue without a distinct convex outer border). This observation about the well-defined and convex outline of a dense breast mass is explained, for example, in the following publications:

- T.F. de Brito Silva, A.C. de Paiva, A.C. Silva, G. Braz Júnior, J.D.S. de Almeida, Classification of breast masses in mammograms using geometric and topological feature maps and shape distribution (Res. Biomed. Eng. (2020). doi:10.1007/s42600-020-00063-x);

- N. Safdarian, M. Hedyezadeh, Detection and Classification of Breast Cancer in Mammography Images Using Pattern Recognition Methods, Multidiscip. Cancer Investig. (2019). doi:10.30699/acadpub.mci.3.4.13.

In practice, the density of a region of interest must exceed a certain specified level for that region of interest to be identified in a morphological image of a particular set.

The persistence of each preliminarily identified region is then assessed using different morphological images. The goal is to morphologically confirm regions of interest that are consistently present under several assumptions regarding the medium through which electromagnetic waves propagate (step E6). Persistence assessment refers to the presence of a region of interest located in three-dimensional (3D) space in the same region across multiple morphological images. Here, it will be assessed whether regions of interest identified through morphological processing are found in the same region across multiple images. This assessment involves, in particular, using criteria such as spatial clustering to associate identified regions of interest with one another.

Thus, persistence of presence makes it possible to morphologically confirm a region of interest, in other words, to associate this region with some physical object, if this region of interest is present in a given proportion of the total number of morphological images.

As noted above, at least two sets of pcfib values are preferably considered, and preferably at least three sets of pcfib values. This is important for the persistence assessment step. Indeed, a region of interest will be considered confirmed if it is present in multiple morphological images. In the case of two sets, for a region of interest to be confirmed, it must be present in both images. In the case of three sets, for a region of interest to be confirmed, it must be present in two out of three images or in three out of three images.

Generally speaking, a region of interest will be considered to be consistently present if it is present in at least some percentage of the morphological images, which percentage should be determined according to the type of region visible.

These regions of interest, confirmed in this way through their persistence, then have a high probability of corresponding to a true lesion or tumor, rather than some artifact, for example. This persistence of presence can be associated with the degree of confidence at the detection level.

In an additional embodiment, the imaging system is considered to comprise a group of measuring probes (K > 1) arranged in a horizontal circle around a cylinder made of a dielectric material. This group of probes, arranged in a circle, is movable along the vertical axis. The organ being examined (the mammary gland) is placed within this cylinder, so that it (the organ) is surrounded by a group of sensors. A dielectric matching medium located in the cylinder where the mammary gland is placed allows the electromagnetic waves emitted by the sensors to penetrate the mammary gland. The vertical positions of the sensor group are predetermined so that there are intervals of constant or variable distance between adjacent positions, and the entire set of positions covers the vertical extent of the mammary gland.

In each of the vertical positions of the probe group, multistatic measurements are performed.

As an example of implementing the data collection stage, we consider L > 1 angular sectors, each containing M > 1 probes from a group (M < K). For example, there are L = 18 angular sectors with M = 6 probes in each sector, and in this case, each sector is offset angularly relative to the adjacent sector by R = 1 probe along the perimeter of a group of K = 18 probes.

In each angular sector of M = 6 probes, each probe sequentially transmits, and for each transmitting probe, the other probes in the same angular sector sequentially receive a signal corresponding to the reflected echo signals coming from obstacles encountered, particularly in the mammary gland. All transmission coefficients measured between the transmitting and receiving probes in the same angular sector are considered to form a multistatic data set. This multistatic data set is repeated for a set of L = 18 angular sectors with M = 6 probes in each. The probe group is then shifted by one interval along the vertical axis, and the multistatic measurements are repeated for different angular sectors.

The array configuration then corresponds to an angular sector at a specific vertical position. Collecting data by angular sector makes it possible to navigate around the breast by making a number of assumptions regarding the pcfib values in each angular sector at each vertical position. This makes it possible to optimally account for the variable and heterogeneous structure of the breast in terms of dielectric properties, which can vary according to different observation positions, and to detect the non-uniform angular characteristics of affected breast areas.

Thus, we benefit from sectorization of the region to be imaged, combined with different assumptions regarding the pcfib parameter to improve detection of regions of interest.

Types of images obtained using the present invention

Figure 3 illustrates a reconstructed microwave radar image for a set of pcfib values such that 10%≤pcfib≤60%. This radar image represents a frontal view of the breast. This radar image displays several pixel regions in terms of intensity.

The purpose of morphological processing is to process this type of radar images to identify areas of interest that could correspond to suspicious areas.

Figures 4a, 4b, 4c, 4d, and 4e illustrate several morphological images of a patient's breast obtained after morphological processing of reconstructed microwave radar images for five sets of pcfib parameter values. These morphological images are presented in the frontal projection of the breast. In this example, the morphological images correspond to five sets of pcfib parameter values:

- fig. 4a: 10%≤pcfib≤60%

- fig. 4b: 20%≤pcfib≤50%

- Fig. 4c: 10%≤pcfib≤20%

- fig. 4d: 30%≤pcfib≤40%

- fig. 4e: 50%≤pcfib≤60%.

According to the radar image shown in Fig. 3, morphological processing allows the identification of only one region of interest. This region of interest is consistently present in the five morphological images indicated and is thus confirmed.

It can be seen that in each of the morphological images the identified region of interest has its own contour, different from the other images, which confirms the fact that the radar signature of the identified object in the microwave range changes depending on the considered sets of pcfib values.

Figs. 5a, 5b, 5c, and 5d illustrate morphological images of the breast of another patient, obtained after morphological processing of the reconstructed microwave radar images, also for five sets of pcfib parameter values. These five sets of values are identical to the sets considered in the case of the previous drawings (of the previous patient). In this example, morphological processing made it possible to identify a single region of interest, which is consistently present in four of the five morphological images. Thus, this region of interest is confirmed. The fifth morphological image, corresponding to the set of values 10%≤pcfib≤20%, is not presented here, since no consistently present region of interest was identified in this image. Unlike the previous example, the identified region of interest has the form of a collection of areas, but remains the same and single connected object according to the previously applied processing operations. This illustrates that the radar signature of the detected object in the microwave range varies significantly depending on the considered sets of pcfib values. This region of interest corresponds to the affected area, which has a distributed, highly uneven shape and a highly heterogeneous texture.

Claims (15)

1. A method for processing medical images of human tissue in a region of a patient's body, and in particular in the mammary gland, using a device (1) for medical imaging, comprising a group of probes emitting/receiving electromagnetic waves in the microwave range, wherein said group contains K > 1 probes separated by gaps from each other, the group of probes has P > 1 different configurations defining emitting probes and receiving probes for one or more positions around the region under consideration, wherein the emitting probes are configured to emit microwave signals in order to irradiate the region of the body, and the receiving probes are configured to receive microwave signals after their scattering and reflection in the region under consideration, in an additional embodiment, the probes can be configured to emit and receive simultaneously, wherein the method comprises the following steps, implemented in a processor module of the device for medical imaging, in which: collect (E1) signals generated in P > 1 antenna group configurations; and for each configuration the method contains the steps of: process (E2) the signals collected for N > 1 sets of Ai > 1 values, 1 ≤ i ≤ N, of a parameter (pcfib) characterizing human tissues through which the signals pass, to obtain elementary microwave radar images; (E3) selecting in each of the N sets an elementary microwave radar image in accordance with at least one image focusing criterion, wherein each of the selected elementary images corresponds to one of the values of the parameter (pcfib) in the set under consideration; for one configuration, one elementary image in the set is selected; In this case, the method for each set contains the following steps: reconstruct (E4), based on the selected elementary radar images, each of the radar image configurations of the patient's body region to reconstruct one three-dimensional (3D) radar image for each set; performing morphological processing (E5) on each reconstructed radar image to obtain a morphological image in which one or more regions of interest, if any, that may constitute the affected area are identified, evaluate (E6) the consistency of the presence of each region of interest across the different morphological images obtained to morphologically confirm the region of interest. 2. The method according to claim 1, wherein at the morphological processing stage (E5) the density of one or more pixel regions of the image is determined, wherein the region of interest is identified if the density associated with it is greater than a threshold value. 3. The method according to claim 1 or 2, wherein at least N=2 sets, preferably at least N=3 sets of Ai > 1 values, 1 ≤ i ≤ N, of the characteristic parameter (pcfib) of the medium through which the signals pass are taken into account. 4. The method according to claim 3, wherein said sets completely or partially overlap each other in terms of ranges of change and/or values. 5. The method according to any one of paragraphs 1-4, in which, at the step of assessing the persistence of presence, the percentage of the presence of the region of interest in the morphological images is determined, and the region of interest is considered confirmed if the percentage exceeds a threshold value. 6. The method according to any one of claims 1 to 5, wherein the group of probes comprises K > 1 probes arranged around the area whose image is to be obtained, said group is configured to move between vertical positions around the area whose image is to be obtained, each configuration is an angular sector of probes, each sector contains M > 1 probes, where M < K, each configuration is shifted in angle by at least one probe relative to another configuration; in each angular sector, each of the probes emits in turn to generate signals collected for the configuration, an elementary radar image is determined for each angular sector in each set of values of the characteristic parameter (pcfib). 7. A medical imaging device comprising a processor module configured to perform the method of any one of claims 1-6.