Methods, devices, equipment, storage media, and procedures for selecting pathological samples

Abstract

This application discloses a method, apparatus, device, storage medium, and program product for selecting pathological samples, relating to the field of artificial intelligence technology. The method includes: acquiring transmitted light images and reflected light images corresponding to multiple pathological samples; determining the pathological regions corresponding to each pathological sample based on the transmitted light images and reflected light images; and selecting a target pathological sample from the multiple pathological samples based on the pathological regions corresponding to each pathological sample. The technical solution provided by this application automatically locates the pathological regions in a pathological sample based on the transmitted light images and reflected light images, avoiding the problems of incomplete and unrealistic pathological region location caused by manual methods, thereby improving the accuracy of pathological region location and thus improving the accuracy of pathological sample selection.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular to a method, apparatus, device, storage medium, and program product for selecting pathological samples. Background Technology

[0002] Postoperative histopathological analysis is the gold standard for tumor diagnosis. The selection of pathological samples (such as tissue blocks cut into appropriate sizes) has a significant impact on postoperative histopathological analysis.

[0003] Taking tumor bed localization as an example, related techniques mainly rely on naked-eye observation and tactile sensation to identify the area corresponding to the tumor bed for selecting pathological samples, and then performing postoperative histopathological analysis based on the selected samples. However, tumor beds often have diverse morphologies and are difficult to define precisely. Relying solely on naked-eye observation and tactile sensation to determine the area corresponding to the tumor bed cannot guarantee the integrity and authenticity of the sample taken from the largest surface of the tumor bed, easily leading to missed tumor beds. This results in low accuracy in the selection of pathological samples, and consequently, low accuracy in pathological analysis. Summary of the Invention

[0004] This application provides a method, apparatus, device, storage medium, and program product for selecting pathological samples, which can improve the accuracy and efficiency of locating pathological areas, thereby improving the accuracy and efficiency of selecting pathological samples. The technical solution is as follows:

[0005] According to one aspect of the embodiments of this application, a method for selecting pathological samples is provided, the method comprising:

[0006] Acquire the transmitted light images and reflected light images corresponding to multiple pathological samples;

[0007] Based on the transmitted light image and reflected light image corresponding to each of the pathological samples, the pathological region corresponding to each of the pathological samples is determined.

[0008] Based on the pathological regions corresponding to each of the pathological samples, a target pathological sample is selected from the plurality of pathological samples.

[0009] According to one aspect of the embodiments of this application, an image acquisition system is provided, the image acquisition system comprising: a transmission light device, a polarization device, and an image capturing device;

[0010] The transmitted light device is used to adjust the image acquisition system to the transmission mode;

[0011] The polarization device is used to adjust the image acquisition system to the reflection mode;

[0012] The image capturing device is used to acquire a transmitted light image corresponding to a pathological sample when the image acquisition system is in the transmission mode, wherein the transmitted light image refers to a transmitted light image under the target dynamic range;

[0013] The image capturing device is also used to acquire a reflected light image corresponding to the pathological sample when the image acquisition system is in the reflection mode, wherein the reflected light image refers to the reflected light image after eliminating the glare effect.

[0014] According to one aspect of the embodiments of this application, a pathological sample selection device is provided, the device comprising:

[0015] The image acquisition module is used to acquire transmitted light images and reflected light images corresponding to the multiple pathological samples, respectively;

[0016] The region determination module is used to determine the pathological region corresponding to each of the pathological samples based on the transmitted light image and reflected light image corresponding to each of the pathological samples respectively;

[0017] The sample selection module is used to select a target pathological sample from the plurality of pathological samples based on the pathological regions corresponding to each of the pathological samples.

[0018] According to one aspect of the embodiments of this application, a computer device is provided, the computer device including a processor and a memory, the memory storing at least one instruction, at least one program, code set or instruction set, the at least one instruction, the at least one program, the code set or instruction set being loaded and executed by the processor to implement the above-described method for selecting pathological samples.

[0019] According to one aspect of the embodiments of this application, a computer-readable storage medium is provided, wherein at least one instruction, at least one program, code set, or instruction set is stored in the storage medium, wherein the at least one instruction, the at least one program, the code set, or the instruction set is loaded and executed by a processor to implement the above-described method for selecting pathological samples.

[0020] According to one aspect of the embodiments of this application, a computer program product or computer program is provided, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the aforementioned method for selecting pathological samples.

[0021] The technical solutions provided in this application embodiment may have the following beneficial effects:

[0022] By using transmitted and reflected light images of pathological samples, pathological regions are automatically located, avoiding the incompleteness and inaccuracies often associated with manual pathological region localization. This improves the accuracy of pathological region localization and consequently enhances the certainty of sample selection. Simultaneously, it avoids the low efficiency of manual pathological region localization, which leads to low sample selection efficiency, thus improving the overall efficiency of pathological region localization and sample selection. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a schematic diagram of the implementation environment of a solution provided in one embodiment of this application;

[0025] Figure 2 This is a schematic diagram of an image acquisition system provided in one embodiment of this application;

[0026] Figure 3 This is a flowchart of a method for selecting pathological samples according to an embodiment of this application;

[0027] Figure 4 This is a schematic diagram of a transmitted light image and a reflected light image provided in one embodiment of this application;

[0028] Figure 5 This is a flowchart of a method for obtaining a pathological region according to an embodiment of this application;

[0029] Figure 6 This is a schematic diagram showing the light transmittance of solid tumors, fibrous tissue, and fat, respectively, according to one embodiment of this application.

[0030] Figure 7 This is a histogram of the translucency of solid tumor, fibrous tissue, and fat, respectively, provided in one embodiment of this application;

[0031] Figure 8 This is a schematic diagram of different tissue compositions at different thicknesses provided in one embodiment of this application;

[0032] Figure 9 This is a curve showing the relationship between the light transmittance and thickness of different tissue components provided in one embodiment of this application;

[0033] Figure 10This is a schematic diagram of the boundary information of a transmitted light image provided in one embodiment of this application;

[0034] Figure 11 This is a schematic diagram of the texture information of a transmitted light image provided in one embodiment of this application;

[0035] Figures 12 to 14 Experimental data from three methods—this application, X-ray, and gross observation—are illustrated exemplarily.

[0036] Figure 15 This is a block diagram of a pathological sample selection device provided in one embodiment of this application;

[0037] Figure 16 This is a block diagram of a pathological sample selection device provided in another embodiment of this application;

[0038] Figure 17 This is a block diagram of a computer device provided in one embodiment of this application. Detailed Implementation

[0039] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0040] Artificial intelligence (AI) is the theory, methods, technology, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to achieve optimal results. In other words, AI is a comprehensive technology within computer science that attempts to understand the essence of intelligence and produce a new kind of intelligent machine that can react in a way similar to human intelligence. AI studies the design principles and implementation methods of various intelligent machines, enabling them to possess the functions of perception, reasoning, and decision-making.

[0041] Artificial intelligence (AI) is a comprehensive discipline encompassing a wide range of fields, including both hardware and software technologies. Fundamental AI technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interactive systems, and mechatronics. AI software technologies primarily include computer vision, speech processing, natural language processing, and machine learning / deep learning.

[0042] Computer vision (CV) is a science that studies how to enable machines to "see." More specifically, it refers to machine vision, which uses cameras and computers to replace human eyes in recognizing and measuring targets, and then performs image processing to create images more suitable for human observation or transmission to instruments. As a scientific discipline, computer vision studies related theories and technologies, attempting to build artificial intelligence systems capable of extracting information from images or multidimensional data. Computer vision technologies typically include image processing, image recognition, image semantic understanding, image retrieval, OCR (Optical Character Recognition), video processing, video semantic understanding, video content / behavior recognition, 3D object reconstruction, 3D technology, virtual reality, augmented reality, simultaneous localization and mapping (SLAM), and common biometric recognition technologies such as facial recognition and fingerprint recognition.

[0043] With the research and advancement of artificial intelligence (AI) technology, AI is being studied and applied in various fields, such as smart homes, smart wearable devices, virtual assistants, smart speakers, smart marketing, autonomous driving, drones, robots, smart healthcare, and smart customer service. It is believed that with the development of technology, AI will be applied in more fields and play an increasingly important role.

[0044] The technical solution provided in this application relates to computer vision technology in artificial intelligence. It acquires brightness value data, texture information, and boundary information of a transmission image, as well as feature information of the transmission image, through computer vision technology. Based on the acquired brightness value data, texture information, boundary information, and feature information of the transmission image, the pathological area is located, and then pathological samples are selected based on the pathological area.

[0045] The method provided in this application can be executed by a computer device, which refers to an electronic device with data computing, processing, and storage capabilities. This computer device can be a terminal such as a PC (Personal Computer), tablet computer, smartphone, wearable device, or intelligent robot; or it can be a server. The server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services.

[0046] The technical solutions provided in this application are applicable to any scenario requiring pathological region localization, such as pathological analysis scenarios, pathological sample selection scenarios, and pathological region localization systems. The technical solutions provided in this application can improve the accuracy and efficiency of pathological region localization, thereby improving the accuracy and efficiency of pathological sample selection.

[0047] In one example, such as Figure 1 As shown, taking a pathological area positioning system as an example, the system may include a terminal 10 and a server 20.

[0048] Terminal 10 can be an electronic device such as a mobile phone, tablet computer, PC, or wearable device. A client application for the target application can be installed on terminal 10. The target application can be a pathology analysis application, a pathology sample selection application, a pathology region location application, etc., and this embodiment does not limit the scope of the application.

[0049] Server 20 can be a standalone physical server, a server cluster or distributed system consisting of multiple physical servers, or a cloud server providing cloud computing services. Server 20 is used to provide background services for the client of the target application in terminal 10. For example, server 20 can be the background server for the aforementioned target application (such as a pathological area localization application).

[0050] Terminal 10 and server 20 can communicate via network 30.

[0051] For example, a user (such as a medical professional or researcher) obtains the transmitted light image and reflected light image of a pathological sample through a client of a target application running on terminal 10, and sends the transmitted light image and reflected light image of the pathological sample to a server. Based on the transmitted light image and reflected light image of the pathological sample, the server locates the pathological region in the pathological sample, and then selects the pathological sample based on the pathological region to obtain the target pathological sample. The server sends the target pathological sample (such as the corresponding number) to the client, and the client displays the target pathological sample to the user.

[0052] Of course, in some other examples, the selection process for the pathological region can also be performed in terminal 10, such as in the client of the target application described above. For example, after acquiring the transmitted light image and reflected light image of the pathological sample, the client directly locates the pathological region in the pathological sample based on the transmitted light image and reflected light image, then selects the pathological sample based on the pathological region to obtain the target pathological sample, and finally displays the target pathological sample to the user.

[0053] Please refer to Figure 2The diagram illustrates an image acquisition system according to an embodiment of this application. The image acquisition system 200 includes: a transmission light device 201, a polarization device 202, and an image capturing device 203.

[0054] Image acquisition system 200 is a system for selecting pathological samples. For example, image acquisition system 200 has image acquisition function and image analysis function. Image acquisition system 200 uses image acquisition function to acquire images corresponding to pathological samples, and then uses image analysis function to select pathological samples, finally outputting the target pathological sample.

[0055] Pathological samples refer to samples used for pathological analysis, such as tissue removed during surgery. Pathological samples can be tissue blocks cut into appropriate volumes. For example, in cancer resection surgery, the tissue corresponding to the tumor area is removed, and then the tissue corresponding to the tumor area is cut into tissue blocks of appropriate volumes; these tissue blocks can serve as pathological samples. Optionally, the target pathological sample can be set according to the user's needs. For example, the target pathological sample can refer to the pathological sample with the largest pathological region size among multiple pathological samples (such as the maximum diameter of the tumor, the maximum surface area of ​​the tumor bed, etc.), or it can refer to the pathological sample with a pathological region size greater than a threshold among multiple pathological samples, or it can be a pre-set number of pathological samples sorted from largest to smallest pathological region size.

[0056] In this embodiment, the pathological region refers to the region corresponding to the diseased tissue (i.e., the lesion region). For example, taking a tumor as an example, the pathological region can refer to the region corresponding to the solid tumor, or the region corresponding to the tumor fibers (i.e., the region corresponding to the tumor bed), or it can include both the region corresponding to the solid tumor and the region corresponding to the tumor fibers. This embodiment does not limit the scope of the pathological region, which can be set according to actual usage requirements.

[0057] Optionally, the image acquisition system 200 realizes the image acquisition function through the transmission light device 201, the polarization device 202 and the image capturing device 203.

[0058] The transmission light device 201 is used to adjust the image acquisition system to the transmission mode; the polarization device 202 is used to adjust the image acquisition system to the reflection mode; the image capturing device 203 is used to acquire the transmission light image corresponding to the pathological sample when the image acquisition system is in the transmission mode, and the transmission light image refers to the transmission light image under the target dynamic range; the image capturing device 203 is also used to acquire the reflection light image corresponding to the pathological sample when the image acquisition system is in the reflection mode, and the reflection light image refers to the reflection light image after eliminating the glare effect.

[0059] Transmission mode refers to the mode used to acquire transmitted light images, and reflection mode refers to the mode used to acquire reflected light images. Transmitted light images are images captured using the principle of light transmission. Reflected light images are images captured using the principle of light reflection. The aforementioned target dynamic range can refer to a range that has been widened, such as HDR (High Dynamic Range), meaning the aforementioned transmitted light image can be an HDR transmitted light image. Glare is used to indicate the presence of unsuitable brightness distribution in the field of view. In this embodiment, glare can be used to indicate reflected light from the surface of a pathological sample that interferes with the localization of the pathological area.

[0060] In one example, the image acquisition system 200 also includes a shooting box, which may be made of black frosted opaque acrylic material, with a transparent area or gap at the top for the image capturing device 203 to take pictures. The image capturing device 203 may be disposed on the top of the shooting box. Exemplarily, the image capturing device 203 may refer to a color camera with autofocus, such as an industrial camera with a resolution of 4000*3000 pixels. Optionally, the image capturing device 203 includes a polarizing filter 205. For example, the lens surface of the camera is equipped with a polarizing filter 205, such as a CPL (Circular Polarizer), and anti-glare shooting can be performed by adjusting the angle of the polarizing filter 205. In this embodiment, the polarizing filter 205 is used in conjunction with the polarizing device 202 to eliminate glare on the pathological sample to obtain the reflected light image corresponding to the pathological sample. Optionally, the angle of the polarizing filter 205 may be pre-adjusted, that is, the angle of the polarizing filter 205 may remain unchanged during the shooting process.

[0061] Optionally, the deflection device 202 can utilize polarization technology to eliminate glare. For example, the polarization device 202 includes a light source device with a polarizing film attached to its surface. This light source device can be an LED (Light Emitting Diode) light source, or any white light source. Exemplarily, the polarization device 202 is disposed on the top inner surface of the imaging chamber. For example, two strip-shaped white LED light sources (e.g., 5 watts) are disposed above the top inner surface of the imaging chamber, and the surfaces of these two strip-shaped white LED light sources are covered with polarizing films having the same polarization direction (e.g., 400-700nm, extinction ratio 97%). Optionally, when only the polarization device 202 is turned on, the image acquisition system 200 is in reflection mode. The aforementioned polarizing mirror 205 can be used in conjunction with this light source device (in the on state) to eliminate glare on the pathological sample, thereby obtaining a reflected light image corresponding to the pathological sample.

[0062] Optionally, the transmitted light device 201 is disposed on the bottom inner surface of the imaging box. The transmitted light device 201 can be an LED light source, as well as any white light source. For example, the transmitted light device 201 is a square white LED light source (e.g., 24 watts). Optionally, when only the transmitted light device 201 is turned on, the image acquisition system 200 is in the transmitted mode.

[0063] In one example, the image acquisition system 200 further includes a computer device 204, through which the image acquisition system 200 implements image acquisition functions. Exemplarily, the computer device 204 is used to acquire transmitted light images and reflected light images corresponding to multiple pathological samples respectively; based on the transmitted light images and reflected light images corresponding to each pathological sample, determine the pathological region corresponding to each pathological sample respectively; and based on the pathological regions corresponding to each pathological sample, select a target pathological sample from the multiple pathological samples.

[0064] For example, computer device 204 acquires the transmitted light image and reflected light image corresponding to the pathological sample. Based on the transmitted light image (such as brightness value data, boundary information, and texture information), it determines the transition region corresponding to the pathological sample. Then, based on the reflected light image, it adjusts the transition region to obtain the pathological region corresponding to the pathological sample. Finally, it selects the pathological sample based on the pathological region. The specific selection process for the pathological sample will be explained in detail below and will not be repeated here.

[0065] Optionally, both the transmission light device 201 and the polarization device 202 can be brightness adjusted via a light source controller. The transmission light device 201 and the polarization device 202 can also receive serial port signals from the computer device 204 to achieve brightness adjustment of the transmission light device 201 and the polarization device 202 (i.e., multiple light source devices), but this embodiment does not limit this.

[0066] Optionally, a transparent acrylic plate can be used as the sample tray 206 to hold pathological samples. A multi-tiered bracket is also provided on the side of the imaging box to allow for adjustment of the imaging distance.

[0067] In this embodiment of the application, the image acquisition system 200 can be simply referred to as HDWIS (High Dynamic Range Dual-model White light imaging system).

[0068] In summary, the technical solution provided in this application automatically locates pathological regions in pathological samples based on transmitted and reflected light images, avoiding the incompleteness and lack of accuracy of pathological samples caused by manual pathological region location. This improves the accuracy of pathological region location and, consequently, the certainty of pathological sample selection. Simultaneously, it avoids the low efficiency of manual pathological region location, which leads to low pathological sample selection efficiency, thus improving the efficiency of pathological region location and, consequently, the efficiency of pathological sample selection.

[0069] Furthermore, by employing an image acquisition system with dual modes (i.e., transmission mode and reflection mode), both transmitted and reflected light images can be acquired, reducing image acquisition complexity and thus improving image acquisition efficiency, which is beneficial for improving the selection efficiency of pathological samples. Moreover, compared to selecting pathological samples using costly radiological images (such as X-ray transmission images), this embodiment only requires acquiring transmitted and reflected light images, achieving both accurate selection of pathological samples and reduced image acquisition costs, thereby lowering the cost of pathological sample selection and improving the practicality and scalability of the image acquisition system.

[0070] In addition, selecting pathological samples based on reflected light images that eliminate the effects of glare can further improve the accuracy of pathological sample selection.

[0071] Please refer to Figure 3 It illustrates a flowchart of a method for selecting pathological samples according to an embodiment of this application. The entity executing each step of the method may be... Figure 1 The computer equipment (i.e., terminal 10 or server 20) in the implementation environment of the scheme shown, or Figure 2 The computer device 204 in the image acquisition system 200 shown can include the following steps (301-303).

[0072] Step 301: Obtain the transmitted light image and reflected light image corresponding to multiple pathological samples respectively.

[0073] A pathological sample refers to a sample used for pathological analysis, such as tissue removed during surgery. A pathological sample can be a tissue block cut to an appropriate volume. A transmitted light image is an image captured using the principle of light transmission. A reflected light image is an image captured using the principle of light reflection. Optionally, a transmitted light image refers to a transmitted light image within the target dynamic range, and a reflected light image refers to a reflected light image after eliminating the influence of glare. The target dynamic range can be HDR, meaning the aforementioned transmitted light image can be an HDR transmitted light image. Glare is used to indicate the presence of an unsuitable brightness distribution in the field of view. In this embodiment, glare can be used to indicate reflected light from the surface of the pathological sample that interferes with the localization of the pathological area. The pathological sample, transmitted light image, and reflected light image are described in the same way as in the above embodiments. Content not described in this embodiment can be referred to in the above embodiments and will not be repeated here.

[0074] In one example, the image acquisition system 200 described above (hereinafter referred to as HDWIS) can acquire transmitted light images and reflected light images corresponding to multiple pathological samples, respectively. Reference Figure 4 For the first pathological sample among multiple pathological samples, one reflected light image 401 and five multiple exposure transmitted light images 402 corresponding to the first pathological sample are acquired using HDWSI. Then, the five multiple exposure transmitted light images 402 are combined using HDWSI to obtain one transmitted light image 403 (i.e., an HDR transmitted light image). Optionally, the number of reflected light images 401, the number of exposed transmitted light images 402, and the exposure time can be set and adjusted according to actual needs. For example, the exposure times of the five multiple exposure transmitted light images 402 can be 1 / 30 second, 1 / 50 second, 1 / 100 second, 1 / 120 second, and 1 / 250 second, respectively. The first pathological sample can refer to any one of the multiple pathological samples.

[0075] For example, the camera lens in HDWSI can be a zoom lens with a focal length of 18-55 mm. The maximum magnification corresponding to the lens can be used for local observation of pathological samples or observation of small pathological samples. (Reference) Figure 4Image 404 is a reflected light image without glare removal, while image 405 is a reflected light image after glare removal. Compared to image 404, which contains a large amount of interference information, image 405 provides more accurate and richer lesion information. Image 406 is a local image of image 405 after optical magnification, and image 407 is a local image of transmitted light image 403 after optical magnification. Image 406 can more clearly represent the texture, boundaries, and features in image 405, and image 407 can more clearly represent the texture, boundaries, and features in transmitted light image 403. Thus, selecting pathological samples based on accurate and rich information about lesions, textures, boundaries, and features is beneficial to improving the accuracy of pathological sample selection.

[0076] Step 302: Based on the transmitted light image and reflected light image corresponding to each pathological sample, determine the pathological region corresponding to each pathological sample.

[0077] The pathological area refers to the area corresponding to the diseased tissue (i.e., the lesion area). The scope of the pathological area is not limited in this embodiment of the application, and it can be set according to actual usage needs.

[0078] In one example, step 302 may also include the following sub-steps:

[0079] Step 302a: For the first pathological sample among multiple pathological samples, determine the candidate region corresponding to the first pathological sample based on the first transmitted light image. The first transmitted light image refers to the transmitted light image corresponding to the first pathological sample, and the candidate region refers to the region in the pathological sample that is suspected to be a pathological region.

[0080] In this embodiment, candidate regions in a pathological sample can be determined based on the brightness value, texture information, and boundary information corresponding to the transmitted light image. The specific process is as follows: The brightness value of the first transmitted light image is calculated to obtain the brightness value data corresponding to the first pathological sample; wherein, the brightness value data is used to characterize the brightness value corresponding to each pixel in the transmitted light image; based on the brightness value data corresponding to the first pathological sample, the transition region corresponding to the first pathological sample is obtained; based on the boundary information and texture information of the first transmitted light image, the transition region corresponding to the first pathological sample is adjusted to obtain the candidate region corresponding to the first pathological sample.

[0081] Brightness values ​​are used to characterize the brightness of a pixel, and brightness value data can be used to ensure the light transmittance of an image. The brightness value of a pixel can be calculated based on the components (such as pixel value, grayscale value, etc.) of the pixel in the three primary color channels (i.e., red, green, and blue).

[0082] Optionally, target pixels whose brightness values ​​belong to the target range in the first transmitted light image can be obtained first; then, based on the target pixels, the transition region corresponding to the first pathological sample can be obtained. The target range can be set and adjusted according to actual usage requirements. For example, different or the same target range can be set for different types of tumors; different target ranges can be set for different pathological regions. This embodiment does not limit this. The transition region can refer to a region composed of target pixels; this region can be a continuous region or multiple independent regions.

[0083] For example, refer to Figure 6 and Figure 7 By calculating the brightness values ​​of the transmitted light images corresponding to 60 pathological samples, the brightness value data corresponding to each of the 60 pathological samples were obtained, that is, the light transmittance of different tissue components (such as solid tumors, fibrous tissue, and fat) in the 60 pathological samples were obtained. Simultaneously, the brightness value data of the 60 pathological samples under the three primary color channels were also obtained for reference.

[0084] refer to Figure 6 Solid tumors, fibrous tissues, and fat exhibit different levels of translucency (i.e., light transmittance) in transmitted light images: 1. For aggregated ductal carcinoma in situ, in the corresponding transmitted light image, the translucency of region 602 corresponding to the solid tumor is less than that of region 601 corresponding to the fibrous tissue, and the translucency of region 601 corresponding to the fibrous tissue is less than that of region 603 corresponding to the fat. 2. For invasive ductal carcinoma, in the corresponding transmitted light image, the translucency of region 605 corresponding to the solid tumor is less than that of region 604 corresponding to the fibrous tissue, and the translucency of region 604 corresponding to the fibrous tissue is less than that of region 606 corresponding to the fat. 3. For small-focal invasive ductal carcinoma, in the corresponding transmitted light image, the translucency of region 607 corresponding to the solid tumor is less than that of region 608 corresponding to the fibrous tissue, and the translucency of region 608 corresponding to the fibrous tissue is less than that of region 609 corresponding to the fat. It can be preliminarily determined that the translucency of the solid tumor < the translucency of the fibrous tissue < the translucency of the fat. Optionally, the regions corresponding to the aforementioned solid tumors, fibrous tissues, and fat can be confirmed based on the corresponding WSI (Whole Slide Image).

[0085] Additionally, refer to Figure 7For brightness values, the order was: solid tumor 701 < fibrous 702 < fat 703. The same order also applied to components in the red channel and the green channel. The blue channel image was too blurry to be of analytical value. Based on brightness value data from 60 pathological samples, the transmittance of the three tissue components, from lowest to highest, was: solid tumor 701 (0.12 + / - 0.03) < fibrous 702 (0.15 + / - 0.04) < fat 703 (0.27 + / - 0.07), with significant differences among them (P < 0.01). Among these, 41 pathological samples contained solid tumor tissue components, and a high percentage (36 / 41) of these samples met the condition that the transmittance of solid tumors was lower than that of fibrous tissues.

[0086] Therefore, the pathological region can be preliminarily located based on the brightness value data corresponding to the pathological sample, thus obtaining the transition region corresponding to the pathological sample. For example, the target range can be set to [0.09, 0.15], or (0.09, 0.15], or [0.09, 0.15), or (0.09, 0.15), which is not limited in this embodiment. The target pixels in the first transmitted light image that belong to the target range form the transition region corresponding to the first pathological sample.

[0087] Optionally, embodiments of this application also obtained experimental data on the light transmittance of different tissue components at different thicknesses. (Reference) Figure 8 and Figure 9 For transmitted light images, in pathological sample group 801 with a thickness of 3 mm, the translucency of solid tumors remained the same: < translucency of fibers < translucency of fat; in pathological sample group 802 with a thickness of 5 mm, the translucency of solid tumors remained the same: < translucency of fibers < translucency of fat; and in pathological sample group 803 with a thickness of 7 mm, the translucency of solid tumors remained the same: < translucency of fibers < translucency of fat. Figure 9 for Figure 8 The curves showing the relationship between the translucency of tissue components and their thickness at different thicknesses are shown in the transmitted light images. It can be seen that at the same thickness, the translucency of solid tumors is still lower than that of fibers, which in turn lowers the translucency of fat. This means that the choice of tissue component thickness does not affect the method for selecting pathological regions provided in this embodiment, thus eliminating the influence of pathological sample thickness and improving the applicability of pathological sample selection.

[0088] In one example, after obtaining the transition region corresponding to the first pathological sample, the edge of the transition region corresponding to the first pathological sample can be adjusted according to the boundary information of the first transmitted light image to obtain the intermediate transition region corresponding to the first pathological sample. The regions in the intermediate transition region corresponding to the first pathological sample whose texture information meets the first condition are deleted to obtain the candidate region corresponding to the first pathological sample.

[0089] Optionally, the pathological sample can be optically magnified using the maximum magnification of the imaging lens in the aforementioned image acquisition system to obtain boundary and texture information of the first transmission image, thereby adjusting the transition region corresponding to the first pathological sample. Alternatively, a neural network can be used to extract boundary and texture information, and then, based on the extracted boundary and texture information, the transition region corresponding to the first pathological sample can be adjusted. Examples include convolutional neural networks and deep neural networks. For instance, a candidate region acquisition model can be constructed using a convolutional neural network, trained on a large number of pathological samples to obtain a trained candidate region acquisition model, and then used to acquire the candidate region corresponding to the first pathological sample.

[0090] The aforementioned boundary information can include the boundary distribution of solid tumors and the boundary distribution of fibrous tissue. For example, refer to... Figure 10 In the transmitted light image 1001, the solid tumor corresponds to a boundary 1005, and the fibrous tissue corresponds to a boundary 1006, indicating a clear / slightly blurred boundary between the solid tumor and the fibrous tissue. Furthermore, in 54% (22 / 41) of the pathological samples, a boundary between the solid tumor and the fibrous tissue can be observed. Therefore, based on the boundary information, the edges of the transition region corresponding to the first pathological sample can be further adjusted to obtain the intermediate transition region corresponding to the first pathological sample. For example, when the pathological region refers to the region corresponding to the tumor, the transition region can be adjusted based on the boundary corresponding to the solid tumor. When the pathological region refers to the region corresponding to the fibrous tissue, the transition region can be adjusted based on the boundaries corresponding to the solid tumor and the fibrous tissue, respectively.

[0091] After obtaining the intermediate transition region corresponding to the first pathological sample, adjustments can be made to this region based on texture information. For example, refer to... Figure 11For region 1101 corresponding to a solid tumor and region 1102 corresponding to fibers in the transmitted light image, region 1102 corresponds to regular and directional texture information, while region 1101 corresponds to texture information with varying degrees of mottled, disordered, or absent texture information. Furthermore, in 50 (50 / 60) pathological samples containing fibrous tissue components, the texture information corresponding to the fibrous tissue components was regular and directional. However, in 41 pathological samples containing tumor tissue components, the texture information corresponding to the tumor tissue components was all mottled, disordered, or absent to varying degrees, i.e., lacking regularity and directionality. Therefore, the first condition mentioned above can refer to regularity and directionality, that is, deleting regions with regular and directional texture information in the intermediate transition region corresponding to the first pathological sample to obtain the candidate regions corresponding to the first pathological sample.

[0092] Step 302b: Adjust the candidate region corresponding to the first pathological sample based on the first reflected light image to obtain the pathological region corresponding to the first pathological sample. The first reflected light image refers to the reflected light image corresponding to the first pathological sample.

[0093] Optionally, the candidate region includes a suspicious region and a credible region. The suspicious region refers to a region that is suspected to be a pathological region, and the credible region refers to a region that is confirmed to be a pathological region.

[0094] Regions that cannot be distinguished by brightness value, boundary information, and texture information can be identified as suspicious regions. For example, referring to the above embodiment, let the target range be [0.09, 0.15]. For regions corresponding to 0.11-0.15, there is a possibility that they correspond to fibers. If texture and boundary information cannot be obtained for certain regions (e.g., the region is particularly small), then it is impossible to further determine whether the region is a pathological region using brightness value, boundary information, and texture information; such regions are suspicious regions. Regions that can be distinguished by brightness value, boundary information, and texture information are then identified as reliable regions.

[0095] For example, refer to Figure 10 In the transmitted light image 1001, since region 1007 cannot be distinguished by brightness value, boundary information, and texture information, region 1007 can be identified as a suspicious region. However, the regions corresponding to the tumor (i.e., the region within boundary 1005) and the regions corresponding to the fibers (i.e., the region within boundary 1006 and outside boundary 1005) can be distinguished by brightness value, boundary information, and texture information, and can therefore be identified as reliable regions. Optionally, this can be confirmed by a full-slide digital scan 1004. The X-ray transmission image 1003, however, cannot reflect accurate boundary information and is not of reference value.

[0096] In one example, the process of obtaining the pathological region can be as follows: based on the feature information of the first reflected light image, the suspicious region corresponding to the first pathological sample is detected; if the suspicious region corresponding to the first pathological sample is a pathological region, the candidate region corresponding to the first pathological sample is determined as the pathological region corresponding to the first pathological sample; if the suspicious region corresponding to the first pathological sample is not a pathological region, the suspicious region is removed from the candidate region corresponding to the first pathological sample to obtain the pathological region corresponding to the first pathological sample.

[0097] Among these, feature information can refer to information such as color, toughness, and thickness distribution. For example, reference... Figure 10 In the reflected light image 1002, since the feature information corresponding to region 1007 is grayish-yellow in color, tough in texture, and uneven in thickness, it can be determined that region 1007 is a mixed region corresponding to fiber and fat. Therefore, region 1007 can be deleted, and the remaining reliable region is the pathological region.

[0098] Step 303: Select the target pathological sample from multiple pathological samples based on the pathological regions corresponding to each pathological sample.

[0099] The target pathological sample can be set according to the user's needs. For example, the target pathological sample can refer to the pathological sample with the largest pathological region size among multiple pathological samples (such as the maximum diameter of the tumor, the maximum surface area of ​​the tumor bed, etc.), or it can refer to the pathological sample with a pathological region size greater than a threshold among multiple pathological samples, or it can be a set number of pathological samples sorted from largest to smallest according to the size of the pathological region.

[0100] For example, the process of obtaining the target pathological sample can be as follows: sort multiple pathological samples in descending order according to the size of the pathological region to obtain a pathological sample sequence; and determine the first set number of pathological samples in the pathological sample sequence as the target pathological sample.

[0101] In summary, the technical solution provided in this application automatically locates pathological regions in pathological samples based on transmitted and reflected light images, avoiding the incompleteness and lack of accuracy of pathological samples caused by manual pathological region location. This improves the accuracy of pathological region location and, consequently, the certainty of pathological sample selection. Simultaneously, it avoids the low efficiency of manual pathological region location, which leads to low pathological sample selection efficiency, thus improving the efficiency of pathological region location and, consequently, the efficiency of pathological sample selection.

[0102] Furthermore, by employing an image acquisition system with dual modes (i.e., transmission mode and reflection mode), both transmitted and reflected light images can be acquired, reducing image acquisition complexity and thus improving image acquisition efficiency, which is beneficial for improving the selection efficiency of pathological samples. Moreover, compared to selecting pathological samples using costly radiological images (such as X-ray transmission images), this embodiment only requires acquiring transmitted and reflected light images, achieving both accurate selection of pathological samples and reduced image acquisition costs, thereby lowering the cost of pathological sample selection and improving the practicality and scalability of the image acquisition system.

[0103] In addition, selecting pathological samples based on reflected light images that eliminate the effects of glare can further improve the accuracy of pathological sample selection.

[0104] In addition, by adopting the technical solution provided in the embodiments of this application, the influence of the thickness of the pathological sample on the selection of the pathological sample can be eliminated, thereby improving the applicability of the selection of pathological samples.

[0105] In one exemplary embodiment, by analyzing the pixels of the outlined region (i.e., the pathological region) and incorporating its location coordinates, the sensitivity and specificity of different methods for fibrous and solid tumors in each pathological sample were analyzed, and a histogram was generated. (Reference) Figure 12 Comparative analysis revealed that HDWIS (the present application) exhibited higher sensitivity compared to X-ray (i.e., radiological imaging methods) and gross examination (i.e., related techniques). The sensitivities of HDWIS, X-ray, and gross examination for fibrosis were 82.7%, 69.2%, and 64.2%, respectively; for solid tumors, the sensitivities were 41.0%, 35.4%, and 32.1%, respectively (P < 0.05). However, there was no significant difference in specificity between HDWIS, X-ray, and gross examination. The specificities of HDWIS, X-ray, and gross examination for fibrosis were 87.3%, 93.4%, and 92.6%, respectively; for solid tumors, the specificities were 71.6%, 64.8%, and 69.3%, respectively.

[0106] Analysis revealed that some pathological samples had thicker sections, causing the WSI (Weighted Incision Image) to not accurately represent the true maximum area of ​​the tumor bed. Using the technical solution provided in this application, it was found that in 12 out of 60 pathological samples, the fibrous area obtained using HDWIS was larger than the corresponding fibrous area in the WSI. Deeply incised pathological samples, and further observation of the restored WSI, revealed new fibrous regions (i.e., tumor bed regions) consistent with those described in this application. For example, refer to... Figure 13In the transmitted light image 1301, compared to the uncut WSI 1303, region 1305 corresponding to the fiber was found, and the uncut WSI 1304 also confirmed the presence of region 1305. However, region 1305 was not observed in the X-ray transmitted image 1302. It is evident that the fiber region obtained by HDWIS is more accurate than that obtained by X-ray.

[0107] refer to Figure 14 After correcting the delineation range, a comparison of 12 pathological samples after deep resection with those before deep resection revealed that HDWIS sensitivity to fibers increased by 3.7% (before deep resection: 90%, after deep resection: 93.7%), while X-ray sensitivity decreased by 14.5% (before deep resection: 70.5%, after deep resection: 56%), and gross observation sensitivity decreased by 16.5% (before deep resection: 69.2%, after deep resection: 53.7%). Therefore, HDWIS showed higher sensitivity than X-ray and gross observation (P < 0.05). After deep resection, HDWIS specificity to fibers increased by 18.4% compared to before deep resection (before deep resection: 74.5%, after deep resection: 92.9%, P < 0.05). The results showed that using HDWIS for pathological sample selection can eliminate the interference of tissue thickness to a certain extent. Compared with the traditional method of gross observation, HDWIS can more accurately identify the distribution range of fibrous and solid tumors, which is beneficial to improving the accuracy of pathological sample selection.

[0108] Furthermore, we can broaden our field of view to non-visible light, such as short-wave infrared (900-2500 nm). Hyperspectral technology can obtain an additional dimension of information on top of spatial information: spectral information. Some studies have verified that short-wave infrared and hyperspectral technologies can be used in tumor resection surgery to help determine the edge location of tumors, while the application of broadband hyperspectral technology has made progress in the detection of breast tumors, and Raman spectroscopy three-dimensional imaging technology has been applied to find and evaluate the surgical margins of breast tumors. Therefore, this application can also combine reflected light images and transmitted light images obtained through a broader spectrum of hyperspectral imaging to more accurately locate pathological areas, thereby more accurately selecting pathological samples.

[0109] In summary, the technical solution provided in this application automatically locates pathological regions in pathological samples based on transmitted and reflected light images, avoiding the incompleteness and lack of accuracy of pathological samples caused by manual pathological region location. This improves the accuracy of pathological region location and, consequently, the certainty of pathological sample selection. Simultaneously, it avoids the low efficiency of manual pathological region location, which leads to low pathological sample selection efficiency, thus improving the efficiency of pathological region location and, consequently, the efficiency of pathological sample selection.

[0110] The following are embodiments of the apparatus described in this application, which can be used to execute the embodiments of the method described in this application. For details not disclosed in the apparatus embodiments of this application, please refer to the embodiments of the method described in this application.

[0111] refer to Figure 15 This diagram illustrates a block diagram of a pathological sample selection device according to an embodiment of this application. The device has the functionality to implement the method example described above; this functionality can be implemented in hardware or by hardware executing corresponding software. The device can be the computer device described above, or it can be installed within a computer device. Figure 15 As shown, the device 1500 includes: an image acquisition module 1501, a region determination module 1502, and a sample selection module 1503.

[0112] The image acquisition module 1501 is used to acquire the transmitted light image and reflected light image corresponding to the multiple pathological samples respectively.

[0113] The region determination module 1502 is used to determine the pathological region corresponding to each of the pathological samples based on the transmitted light image and the reflected light image corresponding to each of the pathological samples.

[0114] The sample selection module 1503 is used to select a target pathological sample from the plurality of pathological samples based on the pathological regions corresponding to each of the pathological samples.

[0115] In one exemplary embodiment, such as Figure 16 As shown, the region determination module 1502 includes: a candidate region acquisition submodule 1502a and a pathological region acquisition submodule 1502b.

[0116] The candidate region acquisition submodule 1502a is used to determine the candidate region corresponding to the first pathological sample among the plurality of pathological samples based on the first transmitted light image, wherein the first transmitted light image refers to the transmitted light image corresponding to the first pathological sample, and the candidate region refers to the region in the pathological sample that is suspected to be the pathological region.

[0117] The pathological region acquisition submodule 1502b is used to adjust the candidate region corresponding to the first pathological sample based on the first reflected light image to obtain the pathological region corresponding to the first pathological sample. The first reflected light image refers to the reflected light image corresponding to the first pathological sample.

[0118] In an exemplary embodiment, the candidate region acquisition submodule 1502a is used for:

[0119] Brightness values ​​are calculated on the first transmitted light image to obtain brightness value data corresponding to the first pathological sample; wherein, the brightness value data is used to characterize the brightness value corresponding to each pixel in the transmitted light image.

[0120] Based on the brightness value data corresponding to the first pathological sample, the transition region corresponding to the first pathological sample is obtained;

[0121] Based on the boundary and texture information of the first transmitted light image, the transition region corresponding to the first pathological sample is adjusted to obtain the candidate region corresponding to the first pathological sample.

[0122] In one exemplary embodiment, the candidate region acquisition submodule 1502a is further configured to:

[0123] Obtain the target pixels in the first transmitted light image whose brightness values ​​belong to the target range;

[0124] Based on the target pixel, the transition region corresponding to the first pathological sample is obtained.

[0125] In one exemplary embodiment, the candidate region acquisition submodule 1502a is further configured to:

[0126] Based on the boundary information of the first transmitted light image, the edge of the transition region corresponding to the first pathological sample is adjusted to obtain the intermediate transition region corresponding to the first pathological sample.

[0127] The regions in the intermediate transition region corresponding to the first pathological sample whose texture information satisfies the first condition are deleted to obtain the candidate region corresponding to the first pathological sample.

[0128] In one exemplary embodiment, the candidate region includes a suspicious region and a reliable region, wherein the suspicious region refers to a region suspected of being the pathological region, and the reliable region refers to a region determined to be the pathological region.

[0129] The pathological region acquisition submodule 1502b is used to detect suspicious regions corresponding to the first pathological sample based on the feature information of the first reflected light image.

[0130] If the suspicious area corresponding to the first pathological sample is the pathological area, the candidate area corresponding to the first pathological sample is determined as the pathological area corresponding to the first pathological sample.

[0131] If the suspicious area corresponding to the first pathological sample is not the pathological area, the suspicious area is removed from the candidate area corresponding to the first pathological sample to obtain the pathological area corresponding to the first pathological sample.

[0132] In one exemplary embodiment, the sample selection module 1503 is configured to:

[0133] Based on the size of the pathological region, the multiple pathological samples are sorted in descending order to obtain a pathological sample sequence;

[0134] The predetermined number of pathological samples in the pathological sample sequence are identified as the target pathological samples.

[0135] In an exemplary embodiment, the transmitted light image refers to the transmitted light image under the target dynamic range, and the reflected light image refers to the reflected light image after eliminating the glare effect.

[0136] In summary, the technical solution provided in this application automatically locates pathological regions in pathological samples based on transmitted and reflected light images, avoiding the incompleteness and lack of accuracy of pathological samples caused by manual pathological region location. This improves the accuracy of pathological region location and, consequently, the certainty of pathological sample selection. Simultaneously, it avoids the low efficiency of manual pathological region location, which leads to low pathological sample selection efficiency, thus improving the efficiency of pathological region location and, consequently, the efficiency of pathological sample selection.

[0137] It should be noted that the apparatus provided in the above embodiments is only illustrated by the division of the above functional modules when implementing its functions. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the apparatus and method embodiments provided in the above embodiments belong to the same concept, and the specific implementation process can be found in the method embodiments, which will not be repeated here.

[0138] Please refer to Figure 17 This diagram illustrates the structural block diagram of a computer device according to an embodiment of this application. This computer device can be used to implement the pathological sample selection method provided in the above embodiments. Specifically, it may include the following:

[0139] In some embodiments, the computer device 1700 includes a central processing unit (such as a CPU, GPU, or FPGA) 1701, a system memory 1704 including RAM (Random-Access Memory) 1702 and ROM (Read-Only Memory) 1703, and a system bus 1705 connecting the system memory 1704 and the central processing unit 1701. The computer device 1700 also includes a basic input / output system 1706 to facilitate information transfer between various devices within the server, and a mass storage device 1707 for storing the operating system 1713, application programs 1714, and other program modules 1715.

[0140] Optionally, the basic input / output system 1706 includes a display 1708 for displaying information and an input device 1709 for user input, such as a mouse or keyboard. Both the display 1708 and the input device 1709 are connected to the central processing unit 1701 via an input / output controller 1710 connected to the system bus 1705. The basic input / output system 1706 may also include an input / output controller 1710 for receiving and processing input from multiple other devices such as a keyboard, mouse, or electronic stylus. Similarly, the input / output controller 1710 also provides output to a display screen, printer, or other types of output devices.

[0141] Optionally, the mass storage device 1707 is connected to the central processing unit 1701 via a mass storage controller (not shown) connected to the system bus 1705. The mass storage device 1707 and its associated computer-readable media provide non-volatile storage for the computer device 1700. That is, the mass storage device 1707 may include computer-readable media (not shown) such as a hard disk or a CD-ROM (Compact Disc Read-Only Memory) drive.

[0142] Without loss of generality, the computer-readable medium may include computer storage media and communication media. Computer storage media include volatile and non-volatile, removable and non-removable media implemented using any method or technology for storing information such as computer-readable instructions, data structures, program modules, or other data. Computer storage media include RAM, ROM, EPROM (Erasable Programmable Read-Only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory), flash memory or other solid-state storage technologies, CD-ROM, DVD (Digital Video Disc) or other optical storage, magnetic tape cassettes, magnetic tape, disk storage, or other magnetic storage devices. Of course, those skilled in the art will recognize that the computer storage medium is not limited to the above-mentioned types. The system memory 1704 and mass storage device 1707 described above can be collectively referred to as memory.

[0143] According to an embodiment of this application, the computer device 1700 can also be connected to a remote computer on a network, such as the Internet. That is, the computer device 1700 can be connected to the network 1712 through the network interface unit 1711 connected to the system bus 1705, or the network interface unit 1711 can be used to connect to other types of networks or remote computer systems (not shown).

[0144] The memory also includes at least one instruction, at least one program, code set, or instruction set, which are stored in the memory and configured to be executed by one or more processors to implement the above-mentioned method for selecting pathological samples.

[0145] In one exemplary embodiment, a computer-readable storage medium is also provided, the storage medium storing at least one instruction, at least one program, code set, or instruction set, the at least one instruction, the at least one program, the code set, or the instruction set being executed by a processor to implement the above-described information processing method.

[0146] Optionally, the computer-readable storage medium may include: ROM (Read-Only Memory), RAM (Random-Access Memory), SSD (Solid State Drives), or optical disc, etc. The random access memory may include ReRAM (Resistance Random Access Memory) and DRAM (Dynamic Random Access Memory).

[0147] In one exemplary embodiment, a computer program product or computer program is also provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, causing the computer device to perform the above-described method for selecting pathological samples.

[0148] It should be understood that "multiple" as used herein refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. Furthermore, the step numbers described herein are merely illustrative of one possible execution order. In some other embodiments, the steps may not be executed in numerical order, such as two steps with different numbers being executed simultaneously, or two steps with different numbers being executed in the reverse order of the illustration. This application does not limit this.

[0149] The above description is merely an exemplary embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for selecting pathological samples, characterized in that, The method includes: Multiple pathological samples are acquired, each corresponding to a transmitted light image and a reflected light image. The transmitted light image is a high dynamic range (HDR) transmitted light image, and the reflected light image is glare-eliminating. The reflected light image is captured by an image capturing device, and the glare is eliminated by a polarizing mirror combined with a polarizing device of the image capturing device. The polarizing device includes a light source device with a polarizing film attached to its surface. For the first pathological sample among the plurality of pathological samples, the brightness value is calculated on the first transmitted light image to obtain the brightness value data corresponding to the first pathological sample; wherein, the first transmitted light image refers to the transmitted light image corresponding to the first pathological sample, and the brightness value data is used to characterize the brightness value corresponding to each pixel in the transmitted light image. Based on the brightness value data corresponding to the first pathological sample, a transition region corresponding to the first pathological sample is obtained, wherein the transition region includes pixels whose brightness values ​​belong to the target range; Based on the boundary and texture information of the first transmitted light image, the transition region corresponding to the first pathological sample is adjusted to obtain the candidate region corresponding to the first pathological sample. The candidate region includes a suspicious region and a reliable region. The suspicious region refers to a region that is suspected to be a pathological region, and the reliable region refers to a region that is determined to be the pathological region. Based on the feature information of the first reflected light image, the suspicious area corresponding to the first pathological sample is detected. The first reflected light image refers to the reflected light image corresponding to the first pathological sample. The feature information includes at least one of the following: color, toughness, and thickness distribution. If the suspicious area corresponding to the first pathological sample is the pathological area, the candidate area corresponding to the first pathological sample is determined as the pathological area corresponding to the first pathological sample. If the suspicious area corresponding to the first pathological sample is not the pathological area, the suspicious area is removed from the candidate area corresponding to the first pathological sample to obtain the pathological area corresponding to the first pathological sample. Based on the pathological regions corresponding to each of the pathological samples, a target pathological sample is selected from the plurality of pathological samples.

2. The method according to claim 1, characterized in that, The step of obtaining the transition region corresponding to the first pathological sample based on the brightness value data of the first pathological sample includes: Obtain the target pixels in the first transmitted light image whose brightness values ​​belong to the target range; Based on the target pixel, the transition region corresponding to the first pathological sample is obtained.

3. The method according to claim 1, characterized in that, The step of adjusting the transition region corresponding to the first pathological sample based on the boundary and texture information of the first transmitted light image to obtain the candidate region corresponding to the first pathological sample includes: Based on the boundary information of the first transmitted light image, the edge of the transition region corresponding to the first pathological sample is adjusted to obtain the intermediate transition region corresponding to the first pathological sample. The regions in the intermediate transition region corresponding to the first pathological sample whose texture information satisfies the first condition are deleted to obtain the candidate region corresponding to the first pathological sample.

4. The method according to claim 1, characterized in that, The step of selecting a target pathological sample from the plurality of pathological samples based on the pathological regions corresponding to each of the pathological samples includes: Based on the size of the pathological region, the multiple pathological samples are sorted in descending order to obtain a pathological sample sequence; The predetermined number of pathological samples in the pathological sample sequence are identified as the target pathological samples.

5. An image acquisition system, characterized in that, The system includes: a light transmission device, a polarization device, an image capturing device, and a computer device; The transmitted light device is used to adjust the image acquisition system to the transmission mode; The polarization device is used to adjust the image acquisition system to the reflection mode; The image capturing device is used to acquire a transmitted light image corresponding to a pathological sample when the image acquisition system is in the transmission mode, and the transmitted light image is an HDR projected light image; The image capturing device is also used to acquire a reflected light image corresponding to the pathological sample when the image acquisition system is in the reflection mode. The reflected light image refers to the reflected light image after eliminating the glare effect. The glare is eliminated by the polarizing mirror of the image capturing device in combination with the polarizing device. The polarizing device includes a light source device with a polarizing film attached to its surface. The computer device is configured to acquire transmitted light images and reflected light images corresponding to multiple pathological samples respectively; for a first pathological sample among the multiple pathological samples, the brightness value of the first transmitted light image is calculated to obtain brightness value data corresponding to the first pathological sample; wherein, the first transmitted light image refers to the transmitted light image corresponding to the first pathological sample, and the brightness value data is used to characterize the brightness value corresponding to each pixel in the transmitted light image; based on the brightness value data corresponding to the first pathological sample, a transition region corresponding to the first pathological sample is acquired, the transition region including pixels whose brightness values ​​belong to the target range; based on the boundary information and texture information of the first transmitted light image, the transition region corresponding to the first pathological sample is adjusted to obtain a candidate region corresponding to the first pathological sample, the candidate region including a suspicious region and a reliable region, the suspicious region... The term "domain" refers to a region suspected of being a pathological region, and "credible region" refers to a region confirmed as the pathological region. Based on the feature information of the first reflected light image, the suspected region corresponding to the first pathological sample is detected. The first reflected light image refers to the reflected light image corresponding to the first pathological sample, and the feature information includes at least one of the following: color, texture, and thickness distribution. If the suspected region corresponding to the first pathological sample is the pathological region, the candidate region corresponding to the first pathological sample is determined as the pathological region corresponding to the first pathological sample. If the suspected region corresponding to the first pathological sample is not the pathological region, the suspected region is removed from the candidate region corresponding to the first pathological sample to obtain the pathological region corresponding to the first pathological sample. Based on the pathological regions corresponding to multiple pathological samples, a target pathological sample is selected from the multiple pathological samples.

6. A device for selecting pathological samples, characterized in that, The device includes: The image acquisition module is used to acquire transmitted light images and reflected light images corresponding to multiple pathological samples respectively. The transmitted light image is a high dynamic range (HDR) transmitted light image, and the reflected light image is glare-eliminating. The reflected light image is captured by the image capturing device, and the glare is eliminated by the polarizing mirror of the image capturing device combined with a polarizing device. The polarizing device includes a light source device with a polarizing film attached to its surface. The region determination module is used to calculate the brightness value of a first transmitted light image for a first pathological sample among the plurality of pathological samples, thereby obtaining brightness value data corresponding to the first pathological sample; wherein, the first transmitted light image refers to the transmitted light image corresponding to the first pathological sample, and the brightness value data is used to characterize the brightness value corresponding to each pixel in the transmitted light image; based on the brightness value data corresponding to the first pathological sample, a transition region corresponding to the first pathological sample is obtained, the transition region including pixels whose brightness values ​​belong to the target range; based on the boundary information and texture information of the first transmitted light image, the transition region corresponding to the first pathological sample is adjusted to obtain a candidate region corresponding to the first pathological sample, the candidate region including a suspicious region and a reliable region. The suspicious area refers to an area suspected to be a pathological area, and the credible area refers to an area confirmed as the pathological area. Based on the feature information of the first reflected light image, the suspicious area corresponding to the first pathological sample is detected. The first reflected light image refers to the reflected light image corresponding to the first pathological sample. The feature information includes at least one of the following: color, toughness, and thickness distribution. If the suspicious area corresponding to the first pathological sample is the pathological area, the candidate area corresponding to the first pathological sample is determined as the pathological area corresponding to the first pathological sample. If the suspicious area corresponding to the first pathological sample is not the pathological area, the suspicious area is removed from the candidate area corresponding to the first pathological sample to obtain the pathological area corresponding to the first pathological sample. The sample selection module is used to select a target pathological sample from the plurality of pathological samples based on the pathological regions corresponding to each of the pathological samples.

7. A computer device, characterized in that, The computer device includes a processor and a memory, the memory storing at least one instruction, which is loaded and executed by the processor to implement the method for selecting pathological samples as described in any one of claims 1 to 4.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores at least one instruction, which is loaded and executed by a processor to implement the method for selecting pathological samples as described in any one of claims 1 to 4.

9. A computer program product, characterized in that, The computer program product includes computer instructions that are executed by a processor to implement the method for selecting pathological samples as described in any one of claims 1 to 4.

Images