AI ultrasonic image comprehensive analysis method for multi-organ examination

Abstract

The invention relates to the technical field of ultrasonic scanning, in particular to an AI ultrasonic image comprehensive analysis method for multi-organ examination, which comprises the following steps: respectively acquiring original ultrasonic images for a plurality of target organs of a patient; performing three-dimensional reconstruction based on the original ultrasonic image and outputting a target section image on the target section; and processing according to the target section image to output measurement data. In order to solve the problems that in the prior art, multiple organs can only be diagnosed through multi-organ ultrasonic examination, and associated features, such as arteries, among the multiple organs are difficult to capture, the method comprises the steps that after original ultrasonic images are collected from the multiple organs respectively, the multiple organs are subjected to ultrasonic examination; three-dimensional reconstruction in a panoramic range is performed based on a cross-linked area of an original ultrasonic image, a target section image on a target is re-intercepted for a to-be-analyzed object such as an artery, and then the target section image is measured to obtain various index parameters.

Description

Technical Field

[0001] The present invention relates to the technical field of ultrasonic scanning, and in particular to an AI ultrasonic image comprehensive analysis method for multi-organ examination. Background Art

[0002] Ultrasound scanning, or medical ultrasound examination, is a medical imaging diagnostic technology based on ultrasound. This technology can use ultrasound to visualize the patient's muscle tissue and internal organs outside the patient's body, including size, structure, and pathological lesions, to assist doctors in making effective diagnoses of the patient's disease. Ultrasound scanning equipment refers to medical equipment that uses the above-mentioned ultrasound scanning technology, which uses ultrasound to image the various tissue parts of the patient to assist in medical diagnosis. In order to achieve a better diagnostic effect on the disease, it is usually necessary to compile a corresponding ultrasound scanning diagnostic process and perform the scanning process according to a specific process. For example, for thyroid scanning, it is usually necessary to design the corresponding scanning direction, sequence, and judgment branches, such as how to deal with the nodule site when scanning.

[0003] In the prior art, in order to achieve a more comprehensive diagnosis process for the patient, multiple parts of the patient are usually scanned. For example, Chinese patent CN202310716182.1 discloses a method, medium and system for adjusting the position of a medical ultrasound detection probe, which belongs to the field of medical ultrasound detection technology. The method includes: obtaining an image of a first area, the first ultrasound detection area is the current detection area in the ultrasound detection process; predicting a second area based on the image of the first area, the second area is the area that needs to be detected in the next step in the ultrasound detection process; calculating the recommended position of the detection probe in the second area, and outputting the recommended position of the detection probe to medical staff. Among them, the step of predicting the second area based on the image of the first area includes: obtaining multiple points of interest in the first area based on the image of the first area, the points of interest are points on the organs in the human body related to the disease or examination during medical ultrasound detection; obtaining the most matching point of interest from multiple points of interest based on the trajectory of the medical staff's operation of the detection probe; and predicting the three-dimensional coordinates of the second area based on the most matching point of interest and the image of the first area.

[0004] However, during the actual implementation process, the inventors found that this type of technical solution usually involves separate inspections of points of interest on local organs, which are then summarized and diagnosed by doctors. It does not capture the correlation information between multiple organs well, and it is difficult to provide a diagnostic basis for some diseases involving multiple organs. Summary of the invention

[0005] In view of the above-mentioned problems existing in the prior art, an AI ultrasound image comprehensive analysis method for multi-organ examination is now provided.

[0006] The specific technical solutions are as follows:

[0007] An AI ultrasound image comprehensive analysis method for multi-organ examination, comprising:

[0008] Step S1: acquiring original ultrasound images for multiple target organs of the patient respectively;

[0009] The target organ is distributed in the chest cavity and abdominal cavity of the patient, and the original ultrasound image includes scanning sections corresponding to the target organ at multiple different depths;

[0010] Step S2: performing three-dimensional reconstruction based on the original ultrasound image and outputting a target section image on the target section;

[0011] Step S3: Processing the target section image to output measurement data.

[0012] On the other hand, the step S1 comprises:

[0013] Step S11: determining scanning parameters of each target organ for the patient respectively;

[0014] The scanning parameters include scanning range, scanning angle and scanning depth;

[0015] Step S12: scanning the target organs respectively according to the scanning parameters to obtain the original echo sequence;

[0016] Step S13: reconstructing the original echo sequence to obtain the original ultrasound images respectively.

[0017] On the other hand, the step S11 comprises:

[0018] Step S111: photographing an optical image of a target area of ​​the patient;

[0019] The target area optical image covers the chest and abdomen of the patient

[0020] Step S112: performing edge extraction on the target area optical image to determine a first aspect ratio of the target area;

[0021] Step S113: generating a scaling ratio according to the first aspect ratio and a second aspect ratio of a pre-configured standard scanning template;

[0022] Step S114: adjusting the standard scanning template according to the scaling ratio to obtain the scanning parameters.

[0023] On the other hand, the step S114 includes:

[0024] Step S1141: adjusting the standard scanning template according to the scaling ratio to obtain the scanning parameters, and performing coordinate mapping on the scanning position in the standard scanning template according to the scaling ratio to obtain a simulated scanning position;

[0025] Step S1142: performing a simulated scan at the simulated scan position using the scan parameters to obtain a simulated section projection;

[0026] Step S1143: determining whether the simulated section projection can cover the target section;

[0027] If yes, output the scanning parameters;

[0028] If not, the scanning parameters are increased step by step according to a preset step size, and then the process returns to step S1142.

[0029] On the other hand, the step S2 comprises:

[0030] Step S21: performing three-dimensional reconstruction on the original ultrasound images to obtain three-dimensional reconstruction areas;

[0031] Step S22: extracting cross-linked regions from the three-dimensional reconstructed regions respectively, and splicing the three-dimensional reconstructed regions according to the cross-linked regions to obtain spliced ​​data;

[0032] Step S23: extracting the target section image on the target section from the stitched data.

[0033] On the other hand, the step S22 comprises:

[0034] Step S221: mapping the three-dimensional reconstruction regions to the same coordinate system for registration;

[0035] Step S222: sampling the three-dimensional reconstruction area to obtain a three-dimensional reconstruction point cloud, and extracting a boundary set from the three-dimensional reconstruction area;

[0036] Step S223: determining the cross-linked area according to the boundary set, extracting cross-linked point clouds in the cross-linked area respectively and merging them;

[0037] Step S223: merging the 3D reconstructed point clouds according to the cross-linked point clouds to form the spliced ​​data.

[0038] On the other hand, when the target section is an arterial target section corresponding to an artery, step S23 includes:

[0039] Step S231: intercepting an orthographic image in an orthographic direction according to the stitched data, and marking the arterial region of the artery on the orthographic image using a prediction model;

[0040] Step S232: determining the section position parameters of the target section according to the arterial region;

[0041] Step S233: capturing the target section image according to the section position parameters.

[0042] On the other hand, the step S231 includes:

[0043] Step S2311: intercepting the front view image in the front view direction according to the stitching data;

[0044] Step S2312: Separating the blood vessel image from the front view image according to the grayscale threshold;

[0045] Step S2313: using the prediction model to mark the arterial region of the artery in the vascular image.

[0046] The above technical solution has the following advantages or beneficial effects:

[0047] In view of the fact that multi-organ ultrasound examination in the prior art can only diagnose multiple organs separately, it is difficult to capture the correlation characteristics between multiple organs, such as problems with arteries. In this embodiment, after collecting original ultrasound images of multiple organs separately, three-dimensional reconstruction is performed within a panoramic range based on the cross-linked areas of the original ultrasound images, and the target section image on the target to be analyzed, such as the artery, is re-captured, and then the target section image is measured to obtain various indicator parameters. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The embodiments of the present invention will be described more fully with reference to the attached drawings, which are provided for illustration and description only and are not intended to limit the scope of the present invention.

[0049] Figure 1 It is an overall schematic diagram of an embodiment of the present invention;

[0050] Figure 2 This is a schematic diagram of step S1 in an embodiment of the present invention;

[0051] Figure 3 This is a schematic diagram of step S11 in an embodiment of the present invention;

[0052] Figure 4 This is a schematic diagram of step S114 in an embodiment of the present invention;

[0053] Figure 5 This is a schematic diagram of step S2 in an embodiment of the present invention;

[0054] Figure 6 This is a schematic diagram of step S22 in an embodiment of the present invention;

[0055] Figure 7 This is a schematic diagram of step S23 in an embodiment of the present invention; Figure 8 Schematic diagram of step S231 in an embodiment of the present invention. DETAILED DESCRIPTION

[0056] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0057] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0058] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments, but they are not intended to limit the present invention.

[0059] The present invention comprises:

[0060] An AI ultrasound image comprehensive analysis method for multi-organ examination, such as Figure 1 As shown, including:

[0061] Step S1: acquiring original ultrasound images for multiple target organs of the patient respectively;

[0062] The target organ is distributed in the chest and abdominal cavities of the patient, and the original ultrasound image includes scanning sections corresponding to the target organ at multiple different depths;

[0063] Step S2: Perform three-dimensional reconstruction based on the original ultrasound image and output a target section image on the target section;

[0064] Step S3: Process the target section image to output measurement data.

[0065] Specifically, in view of the fact that multi-organ ultrasound examination in the prior art can only diagnose multiple organs separately, it is difficult to capture the correlation characteristics between multiple organs, such as problems with arteries. In this embodiment, after collecting original ultrasound images of multiple organs separately, three-dimensional reconstruction is performed within a panoramic range based on the cross-linked areas of the original ultrasound images, and the target section image on the target to be analyzed, such as the artery, is re-captured, and then the target section image is measured to obtain various indicator parameters as measurement data, which are provided to doctors as a basis for subsequent inspection and diagnosis.

[0066] In one embodiment, Figure 2 As shown, step S1 includes:

[0067] Step S11: determining scanning parameters of each target organ for each patient;

[0068] Scanning parameters include scanning range, scanning angle and scanning depth;

[0069] Step S12: Scanning the target organs according to the scanning parameters to obtain original echo sequences;

[0070] Step S13: reconstruct the original echo sequence to obtain original ultrasound images.

[0071] Specifically, in order to achieve better scanning effects and facilitate subsequent overall reconstruction, in this embodiment, scanning parameters for each target organ are determined according to different patients, including scanning range, scanning angle, and scanning depth, so that while examining the target organ, information of the surrounding area can be better collected, and then the target organs are scanned according to the scanning parameters to obtain original echo sequences, and then the original echo sequences are reconstructed to obtain original ultrasound images.

[0072] In one embodiment, Figure 3 As shown, step S11 includes:

[0073] Step S111: taking an optical image of a target area of ​​a patient;

[0074] Optical images of the target area cover the patient's chest and abdomen

[0075] Step S112: performing edge extraction on the target area optical image to determine a first aspect ratio of the target area;

[0076] Step S113: generating a scaling ratio according to the first aspect ratio and a second aspect ratio of a pre-configured standard scanning template;

[0077] Step S114: Adjust the standard scanning template according to the scaling ratio to obtain scanning parameters.

[0078] Specifically, in order to better preserve the surrounding information for the subsequent modeling process while inspecting the target organ, in this embodiment, before starting the scan, an optical image of the target area of ​​the patient is taken from a top-down direction, including the patient's chest and abdomen. Then, a pre-trained segmentation model is used to extract feature points from the optical image of the target area, such as the vertices of the pelvis on both sides and the shoulders. According to these feature points, a rectangular pattern surrounding the patient's torso can be constructed, and a first aspect ratio is calculated for the rectangular pattern.

[0079] Accordingly, before starting the scan, a standard scanning template is pre-constructed. This standard scanning template is based on the original multi-organ ultrasound examination. By expanding the ultrasound scanning range and depth of each organ, the ultrasound images at multiple depths obtained by scanning multiple ultrasound organs have overlapping scanning processes.

[0080] Since the scanning template is constructed based on the sample image, the corresponding patient volume is different from the actual scanned patient. In order to achieve the same scanning effect on the current patient, the scaling ratio should be generated according to the first aspect ratio and the second aspect ratio of the pre-configured standard scanning template, and then the standard scanning template should be adjusted according to the scaling ratio to obtain the scanning parameters to achieve effective coverage of the patient.

[0081] In one embodiment, Figure 4 As shown, step S114 includes:

[0082] Step S1141: adjusting the standard scanning template according to the scaling ratio to obtain scanning parameters, and performing coordinate mapping on the scanning position in the standard scanning template according to the scaling ratio to obtain a simulated scanning position;

[0083] Step S1142: performing simulated scanning at the simulated scanning position using scanning parameters to obtain simulated section projections;

[0084] Step S1143: determining whether the simulated section projection can cover the target section;

[0085] If yes, output the scan parameters;

[0086] If not, the scanning parameters are increased step by step according to the preset step length, and then the process returns to step S1142.

[0087] Specifically, in order to achieve a better scanning effect, in this embodiment, after adjusting the standard scanning template based on the scaling ratio to obtain the scanning parameters, a verification process is also added. Specifically, firstly, the scanning position in the standard scanning template is mapped to obtain a simulated scanning position by the scaling ratio, and then the scanning parameters are used to perform simulated scanning at the simulated scanning position to obtain simulated section projections at different depths on multiple parts, and it is determined whether this series of simulated section projections can cover the target section. If not, the scanning parameters are gradually increased according to the preset step size and then the simulation and judgment are performed again.

[0088] In one embodiment, Figure 5 As shown, step S2 includes:

[0089] Step S21: performing three-dimensional reconstruction on the original ultrasound images to obtain three-dimensional reconstruction areas;

[0090] Step S22: extracting cross-linked regions from the three-dimensional reconstruction region respectively, and splicing the three-dimensional reconstruction region according to the cross-linked regions to obtain spliced ​​data;

[0091] Step S23: extracting a target section image on the target section from the stitched data.

[0092] Specifically, in order to achieve a better extraction effect of the target section image, in this embodiment, the original ultrasound image is first three-dimensionally reconstructed to obtain a three-dimensional reconstructed area. Among them, the original ultrasound image should contain a series of ultrasound images at different depths. By extracting the ultrasound image according to different grayscale thresholds, the organ body and other tissue parts can be segmented. Subsequently, three-dimensional reconstruction is performed for the different tissue areas obtained by segmentation, including edge extraction, contour fitting, etc., to obtain the overall three-dimensional reconstruction area. The three-dimensional reconstruction areas of different tissues may overlap in the same coordinate system.

[0093] Then, all the three-dimensional reconstruction areas obtained by reconstructing the original ultrasound images are mapped to the same coordinate system, and the mapped areas are adjusted according to the ultrasound coordinate origin. Subsequently, the cross-linked areas are extracted from the three-dimensional reconstruction areas, and the three-dimensional reconstruction areas are spliced ​​according to the cross-linked areas to obtain spliced ​​data, and finally the target section image on the target section is extracted from the spliced ​​data.

[0094] In one embodiment, Figure 6 As shown, step S22 includes:

[0095] Step S221: mapping the three-dimensional reconstruction regions to the same coordinate system for registration;

[0096] Step S222: sampling the three-dimensional reconstruction areas to obtain three-dimensional reconstruction point clouds, and extracting boundary sets from the three-dimensional reconstruction areas;

[0097] Step S223: determining a cross-linked area according to the boundary set, extracting cross-linked point clouds in the cross-linked area and merging them;

[0098] Step S224: merging the 3D reconstructed point clouds according to the cross-linked point clouds to form stitched data.

[0099] Specifically, in order to achieve a better stitching effect, in this embodiment, the three-dimensional reconstruction areas of different organs are respectively mapped to the same coordinate system for registration. During the registration process, the organ origin should be adjusted according to the position of the organ origin relative to the trunk origin, so that the organ origin of each organ has a corresponding relative position in the coordinate system. Then, for each group of three-dimensional reconstruction areas, point cloud processing is performed according to the same sampling rate to obtain a three-dimensional reconstructed point cloud; at the same time, a boundary set is extracted for the three-dimensional reconstruction area, and a cross-linked area is determined according to the boundary set. The cross-linked area is the area enclosed by the coordinate points in each pair of boundary sets. According to this part of the cross-linked area, the cross-linked point cloud is extracted from the three-dimensional reconstruction point cloud and merged. On this basis, the three-dimensional reconstruction point cloud is triangularly meshed, and after the connection, it is merged through the cross-linked point cloud to form stitching data.

[0100] In one embodiment, when the target slice is an arterial target slice corresponding to an artery, Figure 7 As shown, step S23 includes:

[0101] Step S231: intercepting an orthographic image in the orthographic direction according to the spliced ​​data, and marking the arterial region of the orthographic image using a prediction model;

[0102] Step S232: determining the section position parameters of the target section according to the arterial region;

[0103] Step S233: capturing the target section image according to the section position parameters.

[0104] Specifically, in order to achieve a better interception effect, in this embodiment, in the scene of artery analysis, after completing the data splicing, the front view image is intercepted in the front view direction according to the spliced ​​data, and the arterial region of the artery is annotated on the front view image using the prediction model, including the plane projection of the arterial region in the three-dimensional coordinate system and the change along the Z axis. Subsequently, the section position parameters of the target section are determined according to the arterial region, including the plane that can be used for analysis, and finally the target section image is intercepted according to the section position parameters.

[0105] In one embodiment, Figure 8 As shown, step S231 includes:

[0106] Step S2311: capturing a frontal image in the frontal direction according to the stitching data;

[0107] Step S2312: Separating the vascular image from the front view image according to the grayscale threshold;

[0108] Step S2313: using the prediction model to mark the arterial region of the artery in the vascular image.

[0109] Specifically, in order to achieve a better extraction effect, in this embodiment, the stitched data is first intercepted in the frontal direction to obtain the frontal image, and then the vascular image is separated from the frontal image according to the grayscale threshold. Finally, the prediction model is used to mark the arterial area of ​​the artery in the vascular image to complete the extraction process.

[0110] After the above-mentioned extraction process, the ultrasound information of the arterial area can be analyzed, such as Doppler blood flow analysis, so as to analyze the intuitive lesions of each organ under the ultrasound image and the related blood supply conditions, so as to achieve a more comprehensive diagnosis process for patients with lesions in multiple organs.

[0111] The above are only preferred embodiments of the present invention, and are not intended to limit the implementation methods and protection scope of the present invention. Those skilled in the art should be aware that all solutions obtained by equivalent substitutions and obvious changes made using the description and illustrations of the present invention should be included in the protection scope of the present invention.

Claims

1. An AI ultrasound image comprehensive analysis method for multi-organ examination, characterized in that: include: Step S1: acquiring original ultrasound images for multiple target organs of the patient respectively; The target organ is distributed in the chest cavity and abdominal cavity of the patient, and the original ultrasound image includes scanning sections corresponding to the target organ at multiple different depths; Step S2: performing three-dimensional reconstruction based on the original ultrasound image and outputting a target section image on the target section; Step S3: Processing the target section image to output measurement data.

2. The AI ​​ultrasound image comprehensive analysis method according to claim 1, characterized in that: The step S1 comprises: Step S11: determining scanning parameters of each target organ for the patient respectively; The scanning parameters include scanning range, scanning angle and scanning depth; Step S12: scanning the target organs respectively according to the scanning parameters to obtain the original echo sequence; Step S13: reconstructing the original echo sequence to obtain the original ultrasound images respectively.

3. The AI ​​ultrasound image comprehensive analysis method according to claim 2, characterized in that: The step S11 comprises: Step S111: photographing an optical image of a target area of ​​the patient; The target area optical image covers the chest and abdomen of the patient Step S112: performing edge extraction on the target area optical image to determine a first aspect ratio of the target area; Step S113: generating a scaling ratio according to the first aspect ratio and a second aspect ratio of a pre-configured standard scanning template; Step S114: adjusting the standard scanning template according to the scaling ratio to obtain the scanning parameters.

4. The AI ​​ultrasound image comprehensive analysis method according to claim 3, characterized in that: The step S114 comprises: Step S1141: adjusting the standard scanning template according to the scaling ratio to obtain the scanning parameters, and performing coordinate mapping on the scanning position in the standard scanning template according to the scaling ratio to obtain a simulated scanning position; Step S1142: performing a simulated scan at the simulated scan position using the scan parameters to obtain a simulated section projection; Step S1143: determining whether the simulated section projection can cover the target section; If so, output the scanning parameters; If not, the scanning parameters are increased step by step according to a preset step size, and then the process returns to step S1142.

5. The AI ​​ultrasound image comprehensive analysis method according to claim 1, characterized in that: The step S2 comprises: Step S21: performing three-dimensional reconstruction on the original ultrasound images to obtain three-dimensional reconstruction areas; Step S22: extracting cross-linked regions from the three-dimensional reconstructed regions respectively, and splicing the three-dimensional reconstructed regions according to the cross-linked regions to obtain spliced ​​data; Step S23: extracting the target section image on the target section from the stitched data.

6. The AI ​​ultrasound image comprehensive analysis method according to claim 5, characterized in that: The step S22 comprises: Step S221: mapping the three-dimensional reconstruction regions to the same coordinate system for registration; Step S222: sampling the three-dimensional reconstruction area to obtain a three-dimensional reconstruction point cloud, and extracting a boundary set from the three-dimensional reconstruction area; Step S223: determining the cross-linked area according to the boundary set, extracting cross-linked point clouds in the cross-linked area respectively and merging them; Step S223: merging the 3D reconstructed point clouds according to the cross-linked point clouds to form the spliced ​​data.

7. The AI ​​ultrasound image comprehensive analysis method according to claim 5, characterized in that: When the target section is an arterial target section corresponding to an artery, step S23 includes: Step S231: intercepting an orthographic image in an orthographic direction according to the stitched data, and marking the arterial region of the artery on the orthographic image using a prediction model; Step S232: determining the section position parameters of the target section according to the arterial region; Step S233: intercepting the target section image according to the section position parameters.

8. The AI ​​ultrasound image comprehensive analysis method according to claim 1, characterized in that: The step S231 includes: Step S2311: intercepting the front view image in the front view direction according to the stitching data; Step S2312: Separating the blood vessel image from the front view image according to the grayscale threshold; Step S2313: using the prediction model to mark the arterial region of the artery in the vascular image.

Images