FR3082975A1 - VIDEO CLIP SELECTOR FOR MEDICAL IMAGING AND MEDICAL DIAGNOSIS - Google Patents

Abstract

A clip selection method includes receiving via an interface with a medical imaging device a selection of a diagnostic procedure and a target portion of a mammalian body. Then, the device acquires video clip image data from the target portion and stores the video clip image data in an image memory. Each video clip of the video clip image data is then processed by image processing to determine a view and a quality of each video clip and a rule is retrieved from a rule base corresponding to the diagnostic procedure and the target party selected. In this regard, the rule specifies a view and required quality of the video clip image data. Finally, the retrieved rule is applied to the video clip image data as a filter to produce a subset of video clip image data satisfying the specified required view and quality and the video clip image data subset is stored in the image memory.

Description

VIDEO CLIP SELECTOR FOR MEDICAL IMAGING AND MEDICAL DIAGNOSIS

BACKGROUND OF THE INVENTION Field of the Invention [0002] The present invention relates to the field of medical imaging and diagnostics and, more particularly, the selection of video clips for their use in medical imaging and in medical diagnosis.

Description of the Prior Art [0004] Medical imaging refers to the method of creating a visual representation of an internal part of a mammalian cotps for the purposes of clinical analysis and intervention. medical. Medical imaging seeks to reveal internal structures hidden outside the cotps in order to facilitate the diagnosis and treatment of a disease. Medical imaging incorporates several different methods for image acquisition. Common modalities include radiological devices such as x-ray radiography including computed tomography (CT), magnetic resonance imaging (MR), medical ultrasound or ultrasound, endoscopy, elastography, tactile imaging, thermography, medical photography and functional imaging techniques in nuclear medicine, such as positron emission tomography (PET) and mono-photonic emission tomography (TESP). Depending on the desired use of imaging for the purpose of medical diagnosis or targeting a specific tissue or a particular organ _or prefer different methods for different imaging.

10005] Medical imaging of a target area of the body can be obtained from many different views. Strictly speaking, since medical imagery can be two-dimensional in nature, the angle and approach of the imaging device will result in a different perspective view of the target area. As with the medial radiation modality, a particular view of the target area presented in a medical image may be preferred depending on the desired use of the radiography for medical diagnosis or mr screening purposes. specific tissue or a circumcircular organ or a part thereof.

10006] Finally, a medical image of a target area of the body can vary in quality. In fact, depending on the operator - usually a technician, not the doctor who ultimately makes a diagnosis based on imagery - the clarity and focus of a medicate image can vary. In certarns case was attempted to a target organ may be incomplete omitting key features of the target organ of vision due to an improper setting _pl ace of the imaging sensor. In other circumstances, external factors such as the anatomical characteristics of the body may prevent some clarity of key characteristics of the target organ despite adequate placement of the imaging sensor.

The traditional workflow for medical imaging begins with the use of the imaging modality by a technician on the patient in order to acquire a set of image data. The image data can be still image data or video clip image data depending on the mode. Generally, the technician is aware of the ultimate goal of image data in order to diagnose a particular disease or dysfunction of a target organ. Once acquired, the image data set is stored in a centralized repository, typically called "PACS" or "Image Archive Communication System" and a report, digital or written, is prepared for review by the doctor. The midwife then retrieves the image data set and the report a little later and performs an analysis of the image data. The analysis generally requires the physician to select the most appropriate images from the image dataset of the correct views and the correct quality.

This method can be relatively tedious - in particular because of the need for the doctor not only to select the correct images of the correct views and of the correct quality, but also to efficiently arrange on the display screen the “Correct swimming to facilitate a diagnosis by the simultaneous review of multiple images of different interest.

To the extent that the required quality of an image does not exist in the image data set, but is necessary, or to the extent that the required view of an image does not exist in the image data set , the doctor must then finger the patient so that he returns to make an additional appointment

P of the technician in order to reacquen the missing image data. This constitutes a huge waste of resources for the patient, the healthcare facility and the doctor.

BRIEF SUMMARY OF THE INVENTION

10009] Aspects of embodiments of the invention summarized herein refer to the aforementioned deficiencies and provide a new and non-manifest method, data control system, and computer program product for clip selection for medical imaging. A method of the invention includes receiving, through an interface with a medical imaging device, a selection of a diagnostic procedure and a target portion of a mammalian emps. Next, the medical imaging device acquires a video clip image data quality of the target position and stores the video clip image data in an image memory. Each video clip of the video clip image data is then processed by image beating to determine a view and a quality of each video clip and a rule is retrieved from a rule base corresponding to the diagnostic procedure and to the selected target part. . In this regard, the rule specifies a required quality and quality of the video clip image data so as to obtain a particular measurement necessary in carrying out the diagnostic procedure. Finally, the retrieved rule is applied to the video clip image data as a filter to produce a subset of video clip image data meeting the specified view and required quality and the video clip image data subset. is stored in the image memory.

In a particular aspect of the embodiment, provided that a video clip satisfying the required view and quality specified in the video clip image data is determined during the application of the retrieved rule as not existing in image data from the video clip, an alert is generated via the interface of the medical imaging device. In another aspect of the embodiment, the image processing of each video clip includes subjecting each video clip to a trained neural network to generate output indicating a recognized view in a video clip subjected to a specified confidence level or , alternatively, subjecting each video clip to a content-based image retrieval system designed to compare a video clip submitted to a memory of known image data of particular views in order to indicate a view recognized in the video clip submitted.

In yet another aspect of the embodiment, the image processing of each video clip further includes the computerized calculation of a generalized degree of the quality of an image submitted based on the specified level of product confidence in the output of the neural network. Finally, in yet another aspect of the embodiment, the rule further specifies a required presence of a tracking characteristic in the view. In this way, image processing of each video clip includes subjecting each video clip to a trained neural network to generate an output indicating a recognized cue characteristic in a video clip subjected to a specified level of confidence so that at during application of the retrieved rule, an absence of the tracking characteristic is correlated with a poor quality of the video clip, but that a presence of the tracking characteristic is correlated with a good quality of the video clip.

In another embodiment of the invention, a medical imaging data processing system is configured for clip selection. The system includes a host computer system that includes one or more computers, each with memory and at least one processor. A diagnostic imaging computer program runs in the memory of the host computer system and provides control instructions to a paired medical imaging device in communication mode. The program also provides an interface with the medical imaging device.

Importantly, the program further includes computer program instructions activated during execution to perform a clip selection process for medical imaging. The method includes receiving via an interface to a medical imaging device, a selection of a diagnostic procedure and a target part of a mammalian body. Then, the medical imaging device acquires a plurality of video clip image data from the target portion and stores the video clip image data in an image memory. Each video clip of the video clip image data is then processed by image processing to determine a view and a quality of each video clip and a rule is retrieved from a rule base corresponding to the diagnostic procedure and the target party selected. In this regard, the rule specifies a required view and quality of the video clip image data. Finally, the retrieved rule is applied to the video clip image data as a filter to produce a subset of video clip image data satisfying the specified view and required quality, and the video clip image data subset is stored in the image memory.

Additional aspects of the invention will be mentioned in part in the description which follows and will be partly apparent from the description or can be learned by practicing the invention. The aspects of the invention will be achieved and achieved by means of the elements and combinations detailed in particular in the claims appended hereto. It is understood that both the above general description and the following detailed description are given only by way of example and explanation and do not provide restrictions to the invention as claimed.

BRIEF DESCRIPTION OF THE VARIOUS VIEWS OF THE DRAWINGS The attached drawings, which are incorporated and form part of the present specification, illustrate embodiments of the invention and, together with the description, serve to explain the principles of the invention. 'invention. The embodiments illustrated here are currently preferred, but it is understood that the invention is not limited to the precise arrangements and instrumentalities illustrated, in which:

Figure 1 is an illustration of a clip selection process for medical imaging;

10017) Figure 2 is a schematic illustration of a data processing system configured for clip selection for medical imaging; and

10018) Figure 3 is a flow diagram illustrating a clip selection process for medical imaging.

DETAILED DESCRIPTION OF THE INVENTION

100 ”) Embodiments of the invention vtsent ia video clip selection for medical device imaging and medical diagnostics. According to an embodiment of the invention, a set of video clips simultaneously acquired from a mammalian organ is selected for processing and each video clip of the set is subjected to a trained neural network to classify each video clip according to a particular view of the organ and a method used to acquire the video clip. The confidence produced by the neural network serves as input to a function determining the quality of the video clip. Intentional use of acquired video dtps is then specified, for example, to calculate an organ measurement based on the calculation of a measurement in relation to a specified diagnostic procedure, and a rule from a retrieved rule base indicating a specific view, modality and quality of the intended use. Optionally, the ruler indicates an arrangement for presenting video clips in a player. On the basis of the indication by the rule of sight, the modality and the quality requirement specific for the intended use, the video clips acquired are filtered in order to produce a subset of video clips of the view, specific terms and quality.

Finally, the subset of video clips is provided as input to a diagnostic reader presenting the subset of video clips for observation by a healthcare professional. Optionally, the reader arranges the presentation of the subset of video clips according to the rule. In particular, the layout of the presentation of the video clip subset may necessitate a rearrangement of the video clip subset so that the most informative video clips are first presented to the healthcare professional.

[0020] In _another illustration, Figure 1 is an illustrator of a method of selecting music fo _r medical imaging. As shown in Figure 1, a medical imaging device 100 acquires image data in a set of images 130 of a target organ of a mammalian body. Each image of the set of images 130 reflects a specific view of the target organ and has a particular quality. A _{circ "it selector logic.}

each of the images in the set of images 130 and determines both a view reflected by the image and also a determined quality 160. Next, an intentional procedure 110 is specified for the medical imaging device 100 and a specific rule 140 as to the quality and the required view of the image data is selected in a rule table 120 on the basis of the procedure intended 110 and, optionally, of a specified measure to be calculated according to the procedure intended 110. The rule 140 is supplied to the chip selector logic 150.

The chip selector logic circuit 150 applies the selected rule

140 at each image of the set of images 130 so as to create a filter 180 filtering or sorting (rearranging) the set of images 130 into a subset of images 190 only images having a view and a quality assigned 160 5 sufficient to support the specified procedure 110 and optionally in order with the most desirable view and quality positioned at the top of the schedule. Then, since the chip selector logic 150 determines that one or more German images is or are present in the image subset 190, the image subset 190 is stored in a memory of data 175 for use in a medical diagnosis of the specified procedure 110. Otherwise, the chip selector logic 150 when detecting an empty set 185 for the subset of images 190 sends an invitation 195 to the device medical imaging 100 indicating a need to reacquire new image data satisfying one and / or the other of the required vision or of a quality required for your specific procedure 110.

10022) The method described in connection with Figure 1 can be implemented in a data processing system. In another illustration, Figure! schematically shows a data processing system configured for clip selection for medical imaging. The system includes a host multimedia system 200 which includes a memory 220, at least one processor 210 and a display 230. The host computer system 200 is also coupled to a medical imaging device 250 which is designed to acquire medical mage data from target organs and to an image memory 240 in which the acquired medical rmage data are stored. A _system operating _me e 260 executes in memory 220 of host infonnatique system 200. The operating system 260 supports exécutton a program code of a clip selection module 300.

The program code of the clip selection module 300 is activated during execution by the processor 210 in the memory 220, to receive an indication of a procedure in an interface with the dtspositlf d'tmagene medical 250 which relates to a target organ, together with a set of images in the memory of images 240 acquired by the medical imaging device 250 with respect to the procedure described. The program code is also activated, during execution, to analyze and assign to each image of the set of images both a view and a quality of each image. In this regard, the program code of the clip selection module 300 can supply each image of the set of images to a neural network! 280 trained to produce a probabilistic mdication of a quality and a view for the supplied image. Alternatively, the program code of the clip selection module 300 can supply each mage to a content-based image retrieval system 270 which is able to compare the image data of each image to a known set of image data in order to classify each image in relation to my particular view and to my particular quality on the basis of a mage database of a known view and of a known quality.

Optionally, the content-based image recovery system 270 may indicate a quality based on an appearance in an image of a portion of a landmark of the mammalian body which should be shown compared to the particular view of the image. The absence of the benchmark indicates poor quality. Varying degrees of benchmark presence indicate varying degrees of quality. For example, a clear presence of a landmark in an image when expected indicates good quality. However, a partial presence of the benchmark indicates poor quality.

As another option, an echo distance can be calculated for each image in keeping with a disparity between a pose of an image acquisition device resulting in the image and an optimal pose. More specifically, a set of training images each annotated by a known pose used to acquire one corresponding to the function images and, optionally, a deviation from an optimal pose known a _{priori for} acquiring _{my function.} highest quality image of the formation image, are copied so that a subsequent image, when compared to the formation images, can result in an identification of a similar pose variation to which reference is made as long as the echo distance. The above can be obtained by content-based image recovery or by a neural network trained with training images to indicate the echo distance. A quality is then assigned to the subsequent image based on a comelated echo distance so that a threshold echo distance indicates poorer quality than a smaller echo distance for the image. higher.

As another option, the neural network 280 can indicate a recognized view in a video clip subject to a specified level of confidence.

Once the program code of the clip selection module 300 has established a view and a quality calculated for each image in the set of images, the program code is also activated to select a particular rule from a rule base calibrated on the indicated procedure and to apply the rule to each image in the set of images. In this aspect, the determined view and quality of each image in the set of images is provided as input to the particular rule in order to determine from which view and quality those which exceed those required by the particular rule. If so, the image is added to a subset of images in the image memory 240. Otherwise, the image is discarded. Once each image in the image set has been processed by the particular rule, the program code of the clip selection module 300 determines whether images persist in the subset in the image memory 240. If not , the clip selection module 300 directs the medical imaging device 250 to generate an alert in the interface indicating a need to acquire additional image data.

In yet another illustration of the operation of the clip selection module 300, Figure 3 is a flow diagram illustrating a method for clip selection for medical imaging. Processing begins at block 310 in which a target procedure and a target organ are specified in an interface with the medical imaging device. In block 320, a rule is located for the specified target procedure and target organ. Then, at block 330, a first image in a set of images is retrieved from the data memory and an assigned quality and view are loaded into the memory. Optionally, the recovery is done in real time during the acquisition of 1 set of images by the medical imaging device. In block 340, the localized rule is applied to the quality and view assigned for determination purposes in decision block 350. In decision block 350, if the first image has sufficient quality and view for the procedure target and 1 target organ specified, the first image is added to a subset in block 360. Otherwise, the process continues at decision block 370.

In decision block 370, if additional images remain to be processed, in block 380, a next image in the set of images is selected for processing and the method is repeated through block 340 with 1 application of the localized rule. Otherwise, in decision block 390, it is then determined whether images exist in the subset. If so, at block 400, the subset is returned for use in a diagnostic analysis of the target procedure. But otherwise, at block 410, an invitation is presented in an interface with the medical imaging device indicating a need to acquire additional image data of the required quality, the required view or both.

The present invention can be implemented in a system, method, computer program product, or any combination thereof. The computer program product may include one or more computer readable storage media having computer readable program instructions to cause a processor to realize aspects of the present invention. The Hstble computer storage medium can be a tangible device that can store and store instructions for use by an instruction execution device. The computer-readable storage medium can be, for example, but not limited to, an electromechanical storage device, a magnetic storage device, an optical storage device, an electromechanical storage device, a semi-storage device conductors or any combination of the above devices.

appropriate [003!] _Program instnretions be downloaded to respective calcm / tratternent devices from a computer-readable storage medium or to an external computer or an extensive storage device via a network. Computer readable program insults can run entirely on the computer of

The user, partly on the user's computer, as an iogy package! stand-alone, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. Aspects of the described herein with reference to illustrations of the present invention block diagrams are flow charts and / or methods, apparatuses (systems) and computer program products according to embodiments of the invention. It is understood that each block of flowchart illustrations and / or block combination diagrams in the illustrations of flowcharts and / or sequenced and of the sequenced diagrams can be implemented by computer readable program instructions.

These computer readable program instructions can be provided to a processor of a universal computer, a special purpose computer or other programmable data processing apparatus to produce a machine so that the instmctions , which run through the computer processor or other programmable data processing device, create means for implementing the functions / acts specified in the flowchart and / or in one or more blocks of diagrams 10 sequenced. These computer-stored program instructions Us can also be stored in a computer-readable storage medium which can direct a computer, a programmable data processor and / or other devices to operate in a particular way so that the computer readable storage medium having instructions stored therein includes an article of manufacture including instructions which implement aspects of the function or act specified in the flowchart and / or in one or more blocks of sequenced diagrams.

The computer readable program instructions can also be loaded on a computer, another programmable data processing device or another device to cause a series of operational steps to be performed on the computer, another device. 1 programmable or other device to produce a computer-implemented method so that the instructions that execute on the computer, another programmable device or another device perform the functions / acts specified in the flowchart and / or in one or more blocks of sequenced diagrams.

The flowchart and sequenced diagrams of the figures illustrate 1 architecture, functionality and operation of possible implementation of systems, methods and computer program products according to various embodiments of the present invention. For this purpose, each block of the flowchart or of the sequenced diagrams can represent a module, a segment or a part of instructions, which includes one or more executable instructions to implement the specified logic function (s). In some alternative implementations, the functions noted in the sequence may occur outside the order annotated in the figures. For example, two sequences shown below can actually be executed substantially simultaneously, or the sequences can sometimes be executed in reverse order depending on the functionality involved. It should also be noted that each block of the block diagrams and / or the flowchart illustrations, and combinations of blocks in the block diagrams and / or the flowchart illustrations can be implemented by hardware-based systems for the purposes Specialists who perform the specified functions or acts and perform combinations of computer hardware and instructions for special purposes.

Finally, the terminology used here is used only to describe particular embodiments and is not intended to limit the invention. As used herein, the singular forms "one", "one" and "the or the" are intended to include plural forms also unless the context clearly indicates otherwise. It is further understood that the terms "includes" and / or "including", when used in this specification, specify the presence of characteristics, whole numbers, steps, operations, elements and / or of the components mentioned, but do not prevent the presence or addition of one or more other characteristics, whole numbers, steps, operations, elements, components and / or their groups.

The structures, materials, acts and corresponding equivalents of all means or steps plus functional elements in the claims below are intended to include any structure, material or act to perform the function in combination with other elements claimed as specifically claimed. The description of the present invention has been presented for purposes of illustration and description, but is not intended to be exhaustive or limited to the invention in the form described. Many modifications and variations will appear to those skilled in the art without departing from the scope and spirit of the invention. The embodiment has been chosen and described in order to better explain the principles of the invention and the practical application and allow other persons skilled in the art to understand the invention for various embodiments with various modifications than those suitable for the particular use intended.

After describing the invention of the present application in detail with reference to embodiments thereof, it will be obvious that modifications and variants are possible without departing from the scope of the invention defined in the appended claims as follows:

Claims (20)

1. Clip selection method for medical imaging comprising: receiving through an interface with a medical imaging device, a selection of a diagnostic procedure and a target part of a mammalian body; 1 acquisition of a plurality of video clip image data of the target part using the medical imaging device; storing the video clip image data in an image memory; image processing of each video clip of the video clip image data to determine a view and quality of each video clip; retrieving a rule from a rule base corresponding to the diagnostic procedure and the selected target part, the rule specifying a required view and quality of the video clip image data; applying the retrieved rule to the video clip image data in the form of a filter to produce a subset of video clip image data satisfying the specified view and required quality; and storing the video clip image data subset in the image memory. 2. Method according to claim 1, further comprising: provided that a video clip satisfying the required view and quality specified in the video clip image data is determined when applying the retrieved rule that it does not exist in the video clip image data, the generation an alert via the interface of the medical imaging device. The method of claim 1, wherein processing images of each video chp includes subjecting each video clip to a trained neural network to generate an output indicating a recognized view in a video clip subject to a specified confidence level . The method of claim 1, wherein processing images of each video clip includes subjecting each video clip to a content-based image recovery system designed to compare a video clip submitted to a data store. known images of particular views so as to indicate a recognized view in the submitted video clip. 5. The method as claimed in claim 3, in which the processing of images of each video clip further comprises the computer calculation of a general degree of quality of an image submitted on the basis of the specified level of confidence produced in the output of the neural network. 6. The method of claim 1, wherein the rule further specifies a required presence of a tracking feature in the view, and wherein the Image processing of each video clip includes subjecting each video clip to a trained neural network to generate an output indicating a recognized cue characteristic in a video clip subjected to a specified confidence level, so that during application of the rule retrieved an absence of the tracking characteristic is correlated with a poor quality of the video clip, but a presence of the tracking characteristic is correlated with a good quality of the video clip. The method of claim 1, wherein the storage occurs during a real-time acquisition of the plurality of video clip image data of the target party using the medical imaging device. 8. Medical imaging data processing system configured for clip selection, the system comprising: a host computer system comprising one or more computers, Each including a memory and at least one processor, a diagnostic imaging computer program executing in the memory of the host computer system, the program providing control instructions to a communication-coupled medical imaging device, the program further providing an interface with the medical imaging device, the program also further comprising computer program instructions activated during execution to perform: receiving via the interface a selection of a diagnostic procedure and a target part of a mammalian body; acquiring a plurality of video clip image data of the target party using the medical imaging device; storing the video clip image data in an image memory; image processing of each video clip of the video clip image data to determine a view and quality of each video clip; retrieving a rule from a rule base corresponding to the diagnostic procedure and to the selected target part, the rule specifying a required view and quality of the video clip image data; applying the retrieved rule to the video clip image data in the form of a filter to produce a subset of video clip image data satisfying the specified view and required quality; and storing the video clip image data subset in the image memory of the host computer system. 9. The system as claimed in claim 8, in which the program instructions are further activated to carry out: provided that a video clip satisfying the required view and quality specified in the video clip image data is determined when applying the retrieved rule that it does not exist in the video clip image data, generating an alert via the interface of the medical imaging device. The system of claim 8, wherein processing images of each video clip comprises subjecting each video clip to a trained neural network to generate an output indicating a recognized view in a video clip subjected to a specified confidence level . The system of claims, wherein processing images of each video clip includes subjecting each video clip to a content-based image recovery system designed to compare a video clip subjected to a data memory of known images of particular views so as to indicate a recognized view in the submitted video clip. The system of claim 10, wherein processing images of each video clip further comprises computerized calculation of a generalized degree of quality of a submitted image based on the specified confidence level produced in the output of the neural network. 13. The system of claim 8, wherein the rule further specifies a required presence of a tracking feature in the view and wherein the processing of images of each video clip includes subjecting each video clip to a trained neural network to generate an output indicating a recognized cue characteristic in a video clip subject to a specified confidence level, so that, during the application of the retrieved rule, an absence of the cue characteristic is correlated with poor quality of the clip but the presence of the cue characteristic is correlated with good quality of the video clip. 14. Computer program product for clip selection for medical imaging, the computer program product comprising a computer readable storage medium having program instructions implemented therein, the program instructions being executable by a device for causing the device to carry out a method including: receiving via an interface with a medical imaging device, a selection of a diagnostic procedure and a target part of a mammalian body; acquiring a plurality of video clip image data of the target party using the medical imaging device; storing the video clip image data in an image memory; image processing of each video clip of the video clip image data to determine a view and quality of each video clip; retrieving a rule from a rule base corresponding to the diagnostic procedure and the selected target part, the rule specifying a required view and quality of the video clip image data; applying the retrieved rule to the video clip image data in the form of a filter to produce a subset of video clip image data satisfying the specified view and required quality; and storing the video clip image data subset in the image memory. 15. Computer program product according to claim 14, in which the method carried out by the device further comprises: provided that a video clip satisfying the required view and quality specified in the video clip image data is determined when applying the retrieved rule that it does not exist in the video clip image data, generating an alert via the interface of the medical imaging device. The computer program product of claim 14, wherein processing images of each video clip comprises subjecting each video clip to a trained neural network to generate output indicating a recognized view in a leveled video clip trust desired 17. The computer program product of claim 14, wherein the image processing of each video clip comprises subjecting each video clip to a content-based image recovery system which is designed to compare a video clip subjected to a memory of image data known from particular views so as to indicate a recognized view in the video clip submitted. 18. The computer program product of claim 16, wherein the processing of images of each video clip further comprises the computerized calculation of a generalized degree of quality of a submitted image based on the specified level of confidence produced in exit from the neural network. 19. The computer program product of claim 14, wherein the specified rule additionally a required presence of a tracking characteristic in the view and wherein the processing of images of each video clip comprises subjecting each video clip to a neural network trained to generate an output indicating a recognized cue characteristic in a video clip subjected to a specified confidence level so that, during the application of the recovered rule, an absence of the cue characteristic is correlated with a quality poor video clip, but a presence of the cue characteristic is correlated with good quality of the video clip. The computer program product of claim 14, wherein the storage occurs during the real-time acquisition of the plurality of video clip image data from the target party using the medical imaging device.