Method for rapidly checking digital pathological section and observing missing tissue by pathological diagnosis system

Abstract

The invention provides a method for quickly checking digital pathological sections and observing missing tissues in a pathological diagnosis system. The method comprises the following steps of: 1, creating a region browsing recognition rule, receiving a slice zoom ratio setting parameter and a recognition action set by a user, and determining a corresponding region browsing recognition rule; wherein parameters are set according to the zoom ratio, and a moving frequency rule of different areas of the digital pathological section is determined; determining a stay duration rule of different areas of the digital pathological section according to the identification action; 2, creating an unbrowsed area reminding rule; receiving a preset region browsing recognition rule, judging whether an unbrowsed region of the digital pathological section is a potential cancer region or not, and performing reminding through a preset reminding mode; wherein the preset reminding mode comprises a strong reminding mode and a weak reminding mode; and step 3, displaying the digital pathological section: rendering an area browsed by the digital pathological section according to the prompt, and displaying an area which is not rendered.

Description

Technical Field

[0001] The present invention relates to the technical field of processing digital pathological sections, and in particular to a method for quickly checking digital pathological sections and observing missed tissues in a pathological diagnosis system. Background Art

[0002] Artificial intelligence prediction of cancer areas in digital pathology sections is one of the hot research directions in the field of digital pathology. Traditional pathology section observation requires pathologists to use a microscope to judge the pathological condition of the tissue, but this method has disadvantages such as subjectivity and time-consuming. The development of digital pathology has made it possible to digitize pathology sections. Digital sections can be automatically analyzed through artificial intelligence technology, thereby reducing the impact of human factors on diagnostic results and improving the accuracy and efficiency of pathological diagnosis.

[0003] The combination of digital pathology and artificial intelligence technology makes it possible to automatically analyze pathological sections, especially in the diagnosis of lesion areas. Therefore, the prediction of cancer areas by artificial intelligence in digital pathology sections is not only a hot research direction in the field of digital pathology, but has also been widely used in clinical practice. The advancement of digital pathology provides doctors with more accurate and faster pathological diagnosis and better treatment effects for patients.

[0004] In the medical field, the observation and diagnosis of tissue sections are very critical. Whether it is the traditional method of observing sections through a microscope or viewing sections on a computer screen in a digital section diagnosis system, doctors need to observe the sections in all directions without blind spots. However, due to the limited visual and cognitive abilities of humans, it is difficult for doctors to avoid ignoring or missing individual areas when observing sections, which may affect the accuracy of the judgment results. Especially when the omitted area is a cancerous area, the impact on the diagnosis of the entire pathological section is even greater.

[0005] Therefore, the present invention proposes a missing section reminder function for the digital pathology section cancer area prediction auxiliary diagnosis system, which can remind doctors of the possible missed section areas through strong reminders, thereby minimizing the omission of pathology sections, especially possible cancerous areas.

[0006] Existing digital pathology slide cancer area prediction auxiliary diagnosis systems generally predict cancer areas through algorithm models, but cannot monitor and identify the following situations:

[0007] 1. Whether all potential cancer areas predicted by the system have been reviewed by doctors.

[0008] 2. Whether the non-cancerous areas predicted by the system are checked by the doctor.

[0009] 3. When doctors miss potential cancer areas predicted by the system, there is no effective way to remind them.

[0010] 4. After the system predicts through the algorithm model, it will mark the potential cancer areas, making it difficult for doctors to carefully examine the areas where the cancer areas have not been identified. Summary of the invention

[0011] The present invention provides a method for a pathology diagnosis system to quickly check digital pathology sections and observe missed tissues, so as to solve the problem that the cancer area is predicted by an algorithm model but it is unable to monitor and identify and determine whether there are any missed areas.

[0012] The present invention provides a method for quickly checking digital pathological sections and observing missed tissues in a pathological diagnosis system, comprising:

[0013] Step 1: Create a zone browsing identification rule:

[0014] Receive the slice zoom ratio setting parameters and recognition actions set by the user, and determine the corresponding area browsing recognition rules; wherein,

[0015] According to the parameters set by the zoom ratio, the rules for the number of moves in different areas of the digital pathology slices are determined;

[0016] According to the recognition action, determine the dwell time rules of different areas of the digital pathology slice;

[0017] Step 2: Create unbrowsed area reminder rules;

[0018] Receive the preset area browsing recognition rules, determine whether the unbrowsed area of ​​the digital pathology slide is a potential cancer area, and remind through the preset reminder mode;

[0019] The preset reminder modes include: strong reminder mode and weak reminder mode;

[0020] Step 3: Display of digital pathology slides:

[0021] It is used to render the browsed area of ​​the digital pathology slice according to the reminder, and display the unrendered area.

[0022] As an optional embodiment of the present application: the area browsing recognition rule includes a matching rule of a slice zoom ratio and a recognition action; wherein,

[0023] The slice zoom ratio can be freely set between 1-100x, and the zoom ratio is proportional to the number of times the digital pathology slice needs to be moved on the display;

[0024] The recognition actions include cursor sliding recognition, screen stay recognition and cursor displacement recognition; among them,

[0025] Cursor sliding identification is that when the cursor passes over the areas corresponding to different areas of the digital pathology slice at a set rate, the response is that the doctor has browsed the digital pathology slice of the corresponding area, and the cursor sliding number is preset;

[0026] Screen dwell recognition is used to set the dwell time of different areas of digital pathology slices on the screen. When the preset dwell time is reached, the response is that the doctor has browsed the slices of the corresponding area. The dwell time is preset;

[0027] Cursor displacement recognition is used to set the recognition accuracy of different areas of digital pathology slices under different zoom factors. When the corresponding displacement accuracy is achieved, the corresponding recognition area is responded to have achieved the target recognition accuracy.

[0028] As an optional embodiment of the present application: in the strong reminder mode:

[0029] The unbrowsed area of ​​the digital pathology slide is located as a potential cancer area predicted by the system, and the unbrowsed area is warned and identified by a strong reminder, prompting the user to view the current area, and generating the first browsing data; wherein,

[0030] According to the first browsing data, it is determined whether all areas of the digital pathology slice have been viewed. After all areas have been viewed, the system switches the slice or issues a pathology diagnosis report.

[0031] As an optional embodiment of the present application: in the weak reminder mode:

[0032] If the unbrowsed area of ​​the digital pathology slide is located as a non-potential cancer area predicted by the system, the unbrowsed area is identified by a weak reminder, prompting the user to view the current area, and generating second browsing data;

[0033] It is determined whether there is a pathological feature in the second browsing data, and if there is a pathological feature, the unbrowsed area of ​​the weak reminder is re-marked as a strong reminder mode.

[0034] As an optional embodiment of the present application: the rendering includes:

[0035] Count the areas browsed in the digital pathology slice, determine whether the corresponding area has been browsed, and mark and cut the corresponding area when the browsing is completed; wherein,

[0036] The browsed area is determined by the first rule and the second rule;

[0037] The first rule is to draw the sliding track of the cursor. When all digital pathology slides covered by the track are considered to have been browsed;

[0038] The second rule is to identify the screen stay. When the screen stay identification time meets the preset time, it is considered that the display has completed browsing the digital pathology slices.

[0039] Determine the digital pathology sections that are marked and cut, and render them one by one.

[0040] As an optional embodiment of the present application: the sliding track includes the following drawing steps:

[0041] The server sets the initial cursor position coordinates and the moving position coordinates at each moment through the display image of the digital pathology slice; wherein,

[0042] The display diagram is configured with a global coordinate system based on coordinate tracking;

[0043] When the cursor coordinate position changes, the server generates a first response, and the first response information is used to indicate that the cursor has moved away from the previous coordinate position;

[0044] After the first response is generated in the server, the cursor moves and generates a second response message, the second response message being used to indicate information that the cursor has entered any coordinate point of the digital pathology slice;

[0045] The server generates a sliding track of the cursor according to the first response information and the second response message.

[0046] As an optional embodiment of the present application: the sliding track includes the following drawing steps and also includes:

[0047] Receive the position point information of the initial position coordinates of the cursor on the digital pathology slice, and write the position point information into the cache queue; wherein,

[0048] The location point information includes the timestamp when the user terminal draws the location point and the location information of the location point;

[0049] A preset time length and a screening mechanism for each preset time length are set, and first target location point information corresponding to a timestamp matching the screening time is screened out from the location point information cached in the cache queue, and trajectory drawing is performed based on the first target location point information;

[0050] If the preset duration of trajectory drawing is insufficient under the screening mechanism, trajectory drawing overlap is performed based on the location point information currently cached in the cache queue and the first target location point information drawn within the last preset duration.

[0051] As an optional embodiment of the present application: the screen stay includes:

[0052] Acquire multiple frames of initial images of different regions of the digital pathology slice displayed on the screen;

[0053] Perform region detection on each initial image to obtain time information of different regions on the screen;

[0054] determining first intersection information and second intersection information of time information and area boundaries of different areas of the digital pathology slice,

[0055] The first cross information is used to indicate the first time when the target area of ​​the digital pathology slice enters the screen display;

[0056] The second intersection point is used to indicate a second time when the target area of ​​the digital pathology slice moves out of the screen display;

[0057] The movement time of the target area between the first time and the second time is obtained, and the movement time is used as the residence time of the target area on the screen.

[0058] As an optional embodiment of the present application: the rendering further includes:

[0059] The server receives the browsing information for the target area sent by the application, determines whether the browsing is completed, and responds to the corresponding rendering command when the browsing is completed;

[0060] According to the rendering instruction, the target area is rendered to obtain a rendering result of the target area;

[0061] During the rendering process of the target area, the overall rendering image is drawn.

[0062] As an optional embodiment of the present application: the rendering command is generated based on a pathological matching mechanism; wherein,

[0063] The pathology matching mechanism is connected with the pathology library, pathology feature library and rendering configuration matrix;

[0064] The pathology matching mechanism matches information of different regions of digital pathology sections through different cancer types in the pathology library and pathology features in the pathology feature library;

[0065] Information matching includes pathological feature matching and cancer type matching;

[0066] According to the pathological feature matching results and the cancer type matching results, determining the rendering values ​​corresponding to the corresponding feature parameters and the cancer type parameters in the rendering configuration matrix;

[0067] According to the rendering value, the rendering color and rendering depth are determined, and a rendering command is generated.

[0068] The beneficial effects of the present invention are:

[0069] Through the digital pathology slide cancer area prediction auxiliary diagnosis system, potential cancer areas are identified and browsed, and strong reminders are given to unbrowsed areas and pathological areas that may have cancer features during the identification process. By identifying all areas of the digital pathology slide, weak reminders are given to unbrowsed areas but pathological areas without cancer areas, thus helping doctors to distinguish the contents of digital pathology slides more easily.

[0070] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description and the accompanying drawings.

[0071] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0073] Figure 1 A step diagram of a method for quickly checking digital pathological sections and observing missed tissues in a pathological diagnosis system in an embodiment of the present invention;

[0074] Figure 2 A specific flow chart of a method for quickly checking digital pathological sections and observing missed tissues in a pathological diagnosis system in an embodiment of the present invention;

[0075] Figure 3 The figure is a flow chart of drawing a sliding track in an embodiment of the present invention. DETAILED DESCRIPTION

[0076] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0077] As attached Figure 1 and attached Figure 2 As shown, the present invention provides a method for quickly checking digital pathological sections and observing missed tissues in a pathological diagnosis system, comprising:

[0078] Step 1: Create a zone browsing identification rule:

[0079] Receive the slice zoom ratio setting parameters and recognition actions set by the user, and determine the corresponding area browsing recognition rules; wherein,

[0080] According to the parameters set by the zoom ratio, the rules for the number of moves in different areas of the digital pathology slices are determined;

[0081] According to the recognition action, determine the dwell time rules of different areas of the digital pathology slice;

[0082] Step 2: Create a reminder rule for unbrowsed areas;

[0083] Receive the preset area browsing recognition rules, determine whether the unbrowsed area of ​​the digital pathology slide is a potential cancer area, and remind through the preset reminder mode;

[0084] The preset reminder modes include: strong reminder mode and weak reminder mode;

[0085] Step 3: Display of digital pathology slides:

[0086] It is used to render the browsed area of ​​the digital pathology slice according to the reminder, and display the unrendered area.

[0087] The principle of the above technical solution is:

[0088] This application is mainly applicable to the field of high-resolution imaging technology and computer-aided diagnosis in pathology-assisted diagnosis systems, and is specifically applied to the process in which doctors scan digital pathology sections using OCR technology and conduct all-round observation of digital pathology sections using a cursor.

[0089] In the first embodiment of the present application, as shown in step 1, before the user uses OCR technology to identify digital pathology sections, the pathology diagnosis system first receives the section zoom ratio parameters and recognition actions set by the user. For example, when the color depth of the section is level A, the section zoom ratio is 200%, and the corresponding number of movement of the pathology diagnosis system in the corresponding area is one-third of the reference movement number, and the stay time is one and a half times the preset stay time. When the movement of the pathology diagnosis system in the corresponding area is translated to the left or to the right to reach the preset reference distance, it indicates that the area browsing recognition action is started; specifically, according to these parameters and actions, the pathology diagnosis system intelligently creates corresponding area browsing recognition rules, and can determine the movement number rules of different areas of the digital pathology section according to the zoom ratio setting parameters, and can also determine the stay time rules of different areas of the digital pathology section according to the recognition action.

[0090] In the second embodiment of the present application, as shown in step 2, according to the regional browsing and recognition rules of step 1, the user starts the OCR technology to perform a quick inspection of the digital pathology section, and in the process of quick inspection, judges the first category of areas that do not meet the regional browsing and recognition rules, and the second category of areas that do not meet the regional browsing and recognition rules and have diagnostic recognition abnormalities; the first category of areas and the second category of areas are regarded as potential cancer areas, and through the generation of the first category of areas, a reminder is given in response to the weak reminder mode, and through the generation of the second category of areas, a reminder is given in response to the strong reminder mode.

[0091] In the third embodiment of the present application, as shown in step 3, according to the reminder mode, when there is a weak reminder mode, the present application renders the area corresponding to the digital pathology slice by the first rendering color corresponding to the reminder mode, and when there is a strong reminder mode, the present application renders the area corresponding to the digital pathology slice by the second selected color corresponding to the strong reminder mode.

[0092] The beneficial effects of the present invention are:

[0093] Through the digital pathology slide cancer area prediction auxiliary diagnosis system, potential cancer areas are identified and browsed, and strong reminders are given to unbrowsed areas and pathological areas that may have cancer features during the identification process. By identifying all areas of the digital pathology slide, weak reminders are given to unbrowsed areas but pathological areas without cancer areas, thus helping doctors to distinguish the contents of digital pathology slides more easily.

[0094] Specifically, the area browsing recognition rule includes a matching rule between a slice zoom ratio and a recognition action; wherein,

[0095] The slice zoom ratio can be freely set between 1-100x, and the zoom ratio is proportional to the number of times the digital pathology slice needs to be moved on the display;

[0096] The recognition actions include cursor sliding recognition, screen stay recognition and cursor displacement recognition; among them,

[0097] Cursor sliding identification is that when the cursor passes over the areas corresponding to different areas of the digital pathology slice at a set rate, the response is that the doctor has browsed the digital pathology slice of the corresponding area, and the cursor sliding number is preset;

[0098] Screen dwell recognition is used to set the dwell time of different areas of digital pathology slices on the screen. When the preset dwell time is reached, the response is that the doctor has browsed the slices of the corresponding area. The dwell time is preset;

[0099] Cursor displacement recognition is used to set the recognition accuracy of different areas of digital pathology slices under different zoom factors. When the corresponding displacement accuracy is achieved, the corresponding recognition area is responded to have achieved the target recognition accuracy.

[0100] The principle of the above technical solution is:

[0101] Exemplarily, during the implementation of step 1 of the present application, there is a possible implementation method for the user of the present application, in which the slice zoom ratio and the recognition action are associated and matched with each other, so that when the zoom ratio is different, the corresponding recognition action is also different. For example, when the zoom ratio is reduced or enlarged, the displacement distance and the number of times the corresponding area moves on the display are different in the corresponding recognition action. When making specific settings, the zoom ratio and the number of times the corresponding area of ​​the digital pathology slice moves are proportionally configured to ensure recognition accuracy, ensure that no observation omissions occur during the AI ​​automatic recognition process of identifying digital pathology slices, and prevent doctors from making observation omissions when observing manually.

[0102] In actual implementation, the corresponding browsing recognition rules can be manually configured for active recognition by doctors instead of AI recognition. In this process,

[0103] The browsing and recognition rules are divided into slice zoom ratio and recognition action, that is, when the doctor completes the set action at the set slice zoom ratio, it is regarded that the doctor has completed browsing the entire area in the display at the current magnification.

[0104] The slice zoom ratio can be freely set by the doctor between 1-100x. The larger the zoom ratio is, the more times the digital pathology slice needs to be moved on the display, which means the doctor has higher requirements for browsing the slices.

[0105] On the contrary, the smaller the zoom ratio is, the fewer times the digital pathology slices need to be moved under the display, and the lower the doctor's browsing requirements for the slices.

[0106] There are three types of recognition actions, but cursor displacement recognition is used for automatic recognition scenarios such as AI recognition. In the doctor-active recognition scenario, the doctor can freely choose to set the recognition action, including cursor sliding recognition and screen stay recognition. Cursor sliding recognition means that when the cursor passes through the corresponding area at a set rate, it is considered that the doctor has browsed the slices of the corresponding area, and the cursor sliding number is set by the doctor; screen stay recognition means that when the doctor stays on the screen for a certain period of time, it is considered that the doctor has browsed the slices of the corresponding area, and the stay time is set by the doctor.

[0107] When the slice zoom ratio and the recognition action in the rule take effect at the same time, the browsing operation on the corresponding area of ​​the slice is considered to be completed.

[0108] The beneficial effects of the above technical solution are:

[0109] The present application can match the zoom ratio with the recognition action of the pathological slice, so as to judge whether there is any browsing omission in the pathological slice, and at the same time ensure the accuracy of browsing. The traditional zoom processing can only make the pathological slice zoom in and out, just for the doctor to have a better browsing experience.

[0110] Specifically, in the strong reminder mode:

[0111] The unbrowsed area of ​​the digital pathology slide is located as a potential cancer area predicted by the system, and the unbrowsed area is warned and identified by a strong reminder, prompting the user to view the current area, and generating the first browsing data; wherein,

[0112] According to the first browsing data, it is determined whether all areas of the digital pathology slice have been viewed. After all areas have been viewed, the system switches the slice or issues a pathology diagnosis report.

[0113] The principle of the above technical solution is:

[0114] In the strong reminder mode, this application indicates that the area of ​​the digital pathology slice that has not been fully browsed not only has browsing requirements that do not meet the pre-set regional browsing recognition rules, but is also detected by the system as having the pathological characteristics of a cancer area. At this time, this area is warned by a strong reminder, and is identified and browsed again to generate the first browsing data. The first browsing data not only needs to indicate whether the potential cancer area has been browsed, but also needs to determine whether this area is a cancer area and generate corresponding pathological data. After all is completed, the system switches the slice and observes other areas of the slice. It can also issue a pathological diagnosis report to determine whether the current potential cancer area is a cancer area, as well as the specific detection data of tissue cells.

[0115] The beneficial effects of the above technical solution are:

[0116] The present application can accurately and repeatedly identify areas that not only fail to meet the detection standards but also have pathological characteristics during the detection process of digital pathology sections, determine whether they are cancer areas, and improve the ability to predict and identify cancer risks during the detection process of pathology sections.

[0117] Specifically, in the weak reminder mode:

[0118] If the unbrowsed area of ​​the digital pathology slide is located as a non-potential cancer area predicted by the system, the unbrowsed area is identified by a weak reminder, prompting the user to view the current area, and generating second browsing data;

[0119] It is determined whether there is a pathological feature in the second browsing data, and if there is a pathological feature, the unbrowsed area of ​​the weak reminder is re-marked as a strong reminder mode.

[0120] The principle of the above technical solution is:

[0121] In this embodiment, if the area that the user has not browsed is a non-potential cancer area predicted by the system, that is, the unbrowsed area is simply not browsed according to the preset rules and there is no pathological feature, the unbrowsed area is warned and identified through a weak reminder, prompting the user to check the current area.

[0122] When all areas have been reviewed, you can switch slices or issue a pathology diagnosis report; however, if pathological features are found after review, they will be re-marked as strong reminder mode for further reminder and identification, making the judgment more accurate.

[0123] The beneficial effects of the above technical solution are:

[0124] For the unbrowsed area with weak reminder, if there is still no pathological feature after identification, the system switches the slice and issues a pathological diagnosis report, and handles it according to the normal process;

[0125] However, if pathological features are present, re-identification through the strong reminder mode can achieve more accurate identification, prevent missed detection, and prevent false detection.

[0126] Specifically, in an optional embodiment of the present application, the rendering includes:

[0127] Count the areas browsed in the digital pathology slice, determine whether the corresponding area has been browsed, and mark and cut the corresponding area when the browsing is completed; wherein,

[0128] The browsed area is determined by the first rule and the second rule;

[0129] The first rule is to draw the sliding track of the cursor. When all digital pathology slides covered by the track are considered to have been browsed;

[0130] The second rule is to identify the screen stay. When the screen stay identification time meets the preset time, it is considered that the display has completed browsing the digital pathology slices.

[0131] Determine the digital pathology sections that are marked and cut, and render them one by one.

[0132] The principle of the above technical solution is:

[0133] In this embodiment, for the areas that have been browsed and the areas that have not been browsed of the digital pathology sections, statistics of the browsed areas will be performed. This statistics is performed by rendering image blocks one by one, so as to determine the distinction between the unbrowsed areas and the browsed areas. When the browsing is completed, the distinction is made, and the rendering colors of the two are different; then, different selective colors can be used to mark and cut the unbrowsed areas and the browsed areas.

[0134] In the process of counting the unbrowsed areas and browsed areas, the present application uses two rules to judge whether the browsing or not browsing is done. In the specific judgment process, because the present application sets a cursor, it is possible to judge whether the cursor track covers all the image blocks by the cursor sliding track. If all the image blocks are covered, it means that the corresponding area of ​​the digital pathology slice has been browsed. The second rule is to prevent the first rule from making recognition errors when judging whether to browse, so the first rule and the second rule are used in combination to achieve the goal of no recognition errors. The second rule is the screen dwell time, which can determine whether the screen dwell time meets the preset time, thereby indicating that the corresponding area has completed the browsing, and thus the digital pathology slices after the marked cutting are rendered one by one to determine the unbrowsed pathology area.

[0135] The beneficial effects of the above technical solution are:

[0136] When distinguishing digital pathology sections, the present application can make simultaneous judgments through the first rule and the second rule. While making judgments through these two methods, the cutting lines of the browsed area and the unbrowsed area can be divided through the sliding trajectory, and the unbrowsed area can also be identified. The screen stay recognition can prevent errors in the judgment of the sliding trajectory. At the same time, the judgment rules of the sliding trajectory can also prevent errors in the screen stay recognition.

[0137] Specifically, Figure 3 As shown, in an optional embodiment of the present application, the sliding track includes the following drawing steps:

[0138] The server sets the initial cursor position coordinates and the moving position coordinates at each moment through the display image of the digital pathology slice; wherein,

[0139] The display diagram is configured with a global coordinate system based on coordinate tracking;

[0140] When the cursor coordinate position changes, the server generates a first response, and the first response information is used to indicate that the cursor has moved away from the previous coordinate position;

[0141] After the first response is generated in the server, the cursor moves and generates a second response message, the second response message being used to indicate information that the cursor has entered any coordinate point of the digital pathology slice;

[0142] The server generates a sliding track of the cursor according to the first response information and the second response message.

[0143] The principle of the above technical solution is:

[0144] In the specific implementation process of this application, the server will build a global coordinate system based on the digital pathology slice, and the global coordinate system will be set in the display image of the digital pathology slice. Then, by recording the initial position coordinates of the cursor and the moving position coordinates at each moment on the display image, the browsing record on the digital pathology slice is determined;

[0145] Then, when the cursor position changes, the server generates an initial response signal to record the initial coordinate position. The cursor then moves to the next position, and a second response message is generated at the next position. The second response message indicates the response information for entering the next coordinate position. By repeatedly generating the first response message and the second response message, accurate recording of the cursor movement trajectory is ensured.

[0146] The beneficial effects of the above technical solution are:

[0147] The present application can record the sliding trajectory through the response signal generated in the global coordinate system during the drawing of the sliding trajectory, so as to accurately judge the browsing area. At the same time, when drawing the sliding trajectory through the response information of the cursor, it can also be judged when browsing, based on the distance information of the coordinate points on the global coordinate system between two adjacent responses, whether the browsing is equidistant browsing. Therefore, when performing automatic browsing, the specific performance of AI in the pathological diagnosis system can be judged. If it is not equidistant browsing, it means that the AI ​​performance is not good.

[0148] Specifically, in an optional embodiment of the present application, the sliding track includes the following drawing steps and further includes:

[0149] Receive the position point information of the initial position coordinates of the cursor on the digital pathology slice, and write the position point information into the cache queue; wherein,

[0150] The location point information includes the timestamp when the user terminal draws the location point and the location information of the location point;

[0151] A preset time length and a screening mechanism for each preset time length are set, and first target location point information corresponding to a timestamp matching the screening time is screened out from the location point information cached in the cache queue, and trajectory drawing is performed based on the first target location point information;

[0152] If the preset duration of trajectory drawing is insufficient under the screening mechanism, trajectory drawing overlap is performed based on the location point information currently cached in the cache queue and the first target location point information drawn within the last preset duration.

[0153] The principle of the above technical solution is:

[0154] In the process of recording the sliding trajectory, the present application can generate a corresponding cache queue for the position point information of the cursor in the global coordinate system; the cache queue contains the time and specific coordinate position information of the cursor at each position point. In the process of drawing the trajectory, the sliding trajectory is drawn based on the preset duration (the time period for drawing the sliding trajectory at one end). In this process, the time is filtered (the filtering time is the drawing time point corresponding to the first position point of the next preset time period after the drawing of a preset time period is completed), and then the first target position point information within a preset time period is determined as the initial point for drawing the trajectory within a period of time.

[0155] In the present application, there is also an edge in the digital pathology slice when drawing. When the sliding trajectory approaches the edge, the time period corresponding to the preset time length does not reach the preset time length. At this time, when drawing the trajectory, the trajectory drawing in the previous time period and the current time period are overlapped to generate a longer sliding trajectory exceeding the preset time length, so as to ensure that the drawing trajectory will not be wrong. Because there is a preset time length, the coordinate information at the end of each preset time length is consistent with the next preset time length, that is, the initial coordinate information in the next cursor sliding browsing cycle, thereby ensuring that when drawing the trajectory, the sliding trajectories in countless time periods overlap, and adjacent sliding trajectories can be associated and verified, thereby preventing measurement errors of the sliding trajectory.

[0156] The beneficial effects of the above technical solution are:

[0157] In the process of drawing the sliding trajectory, the screening, matching and trajectory drawing of the cursor position under the preset time length can be used to determine whether there is an error in the trajectory drawing, and ensure that the trajectory drawing is correct. For the trajectory drawing of the edge area of ​​the digital pathology section, the drawing can be extended, so that the drawn trajectory will not be easily ignored because the time length is too short.

[0158] Specifically, in an optional embodiment of the present application, the screen stay includes:

[0159] Acquire multiple frames of initial images of different regions of the digital pathology slice displayed on the screen;

[0160] Perform region detection on each initial image to obtain time information of different regions on the screen;

[0161] determining first intersection information and second intersection information of time information and area boundaries of different areas of the digital pathology slice,

[0162] The first cross information is used to indicate the first time when the target area of ​​the digital pathology slice enters the screen display;

[0163] The second intersection point is used to indicate a second time when the target area of ​​the digital pathology slice moves out of the screen display;

[0164] The movement time of the target area between the first time and the second time is obtained, and the movement time is used as the residence time of the target area on the screen.

[0165] The principle of the above technical solution is:

[0166] In the process of detecting screen stay, the present application will generate multiple frames of initial images according to different regions of the digital pathology slices. Each initial image will perform regional detection to determine the dwell time of the cursor in this region when the user browses. Then, based on the initial image, the different regional boundaries of the digital pathology slices can be determined during the browsing process. The first cross information and the second cross information relative to the browsing time are used to determine the entry time of the cursor into different regions of the digital pathology slices during the browsing process. The second intersection is used to determine the time when the cursor moves out of each region. The first time and the second time can be used to determine the dwell time of the cursor in each region, thereby determining whether the user or AI has browsed each region. Accurate browsing recognition is performed.

[0167] The beneficial effects of the above technical solution are:

[0168] The present application can perform browsing recognition by means of screen dwelling, and judge whether different areas of the digital pathology slice are browsed by means of time information of the cursor controlled by AI or the user entering different areas of the digital pathology slice.

[0169] Specifically, in an optional embodiment of the present application, the rendering further includes:

[0170] The server receives the browsing information for the target area sent by the application, determines whether the browsing is completed, and responds to the corresponding rendering command when the browsing is completed;

[0171] According to the rendering instruction, the target area is rendered to obtain a rendering result of the target area;

[0172] During the rendering process of the target area, the overall rendering image is drawn.

[0173] The principle of the above technical solution is:

[0174] The present invention controls the overall rendering of the browsed area through a server, and the unrendered area is displayed through a screen or a terminal device. In the rendering process, the specific rendering features of the browsed or unbrowsed area can be determined through a rendering command, including the rendering type, the rendering color and the rendering information corresponding to the rendering color.

[0175] The beneficial effects of the above technical solution are:

[0176] The present application can identify browsed areas and unbrowsed areas during the overall rendering process, thereby ensuring that different areas of the digital pathology slice can be distinguished by color during the rendering process.

[0177] Specifically, in an optional embodiment of the present application, the rendering command is generated based on a pathology matching mechanism; wherein,

[0178] The pathology matching mechanism is connected with the pathology library, pathology feature library and rendering configuration matrix;

[0179] The pathology matching mechanism matches information of different regions of digital pathology sections through different cancer types in the pathology library and pathology features in the pathology feature library;

[0180] Information matching includes pathological feature matching and cancer type matching;

[0181] According to the pathological feature matching results and the cancer type matching results, determining the rendering values ​​corresponding to the corresponding feature parameters and the cancer type parameters in the rendering configuration matrix;

[0182] According to the rendering value, the rendering color and rendering depth are determined, and a rendering command is generated.

[0183] The principle of the above technical solution is:

[0184] In the process of generating rendering commands, this application will determine the pathological features of the browsing area and the browsed area through the pathological matching mechanism, match the pathological features with the pathological features in the pathological feature library, and then match the corresponding cancer types in the pathological library through the matching of pathological features. Finally, through the cancer types and the rendering configuration matrix, the rendering values ​​corresponding to the different cancer types and pathological features are determined, and the corresponding rendering commands are generated through the rendering values ​​and color depth. The rendering configuration matrix corresponds to the cancer types and the corresponding pathological feature parameters and types. When the depth of the pathological features is different and the distribution area is different, the rendering values ​​are different.

[0185] The beneficial effects of the above technical solution are:

[0186] This application can use rendering commands corresponding to rendering colors and rendering depths to allow doctors to render areas that may have cancer pathological features with different rendering colors during pathological diagnosis, making it easier to distinguish. Because this application has unbrowsed areas, which have no pathological features, rendering is not performed, but after the unbrowsed areas are re-identified, rendering will be performed again. For areas without pathological features, rendering will not be performed either.

[0187] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for quickly checking digital pathological sections and observing missed tissues in a pathological diagnosis system, characterized in that: include: Step 1: Create a zone browsing identification rule: Receive the slice zoom ratio setting parameters and recognition actions set by the user, and determine the corresponding area browsing recognition rules; wherein, According to the parameters set by the zoom ratio, the rules for the number of moves in different areas of the digital pathology slices are determined; According to the recognition action, determine the dwell time rules of different areas of the digital pathology slice; Step 2: Create a reminder rule for unbrowsed areas; Receive the preset area browsing recognition rules, determine whether the unbrowsed area of ​​the digital pathology slide is a potential cancer area, and remind through the preset reminder mode; The preset reminder modes include: strong reminder mode and weak reminder mode; Step 3: Display of digital pathology slides: It is used to render the browsed area of ​​the digital pathology slice according to the reminder, and display the unrendered area.

2. The method for quickly checking digital pathological sections and observing missed tissues in a pathological diagnosis system according to claim 1, characterized in that: The area browsing recognition rule includes the matching rules of the slice zoom ratio and the recognition action; wherein, The slice zoom ratio can be freely set between 1-100x, and the zoom ratio is proportional to the number of times the digital pathology slice needs to be moved on the display; The recognition actions include cursor sliding recognition, screen stay recognition and cursor displacement recognition; among them, Cursor sliding identification is that when the cursor passes over the areas corresponding to different areas of the digital pathology slice at a set rate, the response is that the doctor has browsed the digital pathology slice of the corresponding area, and the cursor sliding number is preset; Screen dwell recognition is used to set the dwell time of different areas of digital pathology slices on the screen. When the preset dwell time is reached, the response is that the doctor has browsed the slices of the corresponding area. The dwell time is preset; Cursor displacement recognition is used to set the recognition accuracy of different areas of digital pathology slices at different zoom ratios. When the corresponding displacement accuracy is achieved, the corresponding recognition area is responded to have achieved the target recognition accuracy.

3. The method for quickly checking digital pathological sections and observing missed tissues in a pathological diagnosis system according to claim 1, characterized in that: In the strong reminder mode: The unbrowsed area of ​​the digital pathology slide is located as a potential cancer area predicted by the system, and the unbrowsed area is warned and identified by a strong reminder, prompting the user to view the current area, and generating the first browsing data; wherein, According to the first browsing data, it is determined whether all areas of the digital pathology slice have been viewed. After all areas have been viewed, the system switches the slice or issues a pathology diagnosis report.

4. The method for quickly checking digital pathological sections and observing missed tissues in a pathological diagnosis system according to claim 1, characterized in that: In the weak reminder mode: If the unbrowsed area of ​​the digital pathology slide is located as a non-potential cancer area predicted by the system, the unbrowsed area is identified by a weak reminder, prompting the user to view the current area, and generating second browsing data; It is determined whether there is a pathological feature in the second browsing data, and if there is a pathological feature, the unbrowsed area of ​​the weak reminder is re-marked as a strong reminder mode.

5. The method for quickly checking digital pathological sections and observing missed tissues in a pathological diagnosis system according to claim 2, characterized in that: The rendering includes: Count the areas browsed in the digital pathology slice, determine whether the corresponding area has been browsed, and mark and cut the corresponding area when the browsing is completed; wherein, The browsed area is determined by the first rule and the second rule; The first rule is to draw the sliding track of the cursor. When all digital pathology slides covered by the track are considered to have been browsed; The second rule is to identify the screen stay. When the screen stay identification time meets the preset time, it is considered that the display has completed browsing the digital pathology slices. Determine the digital pathology sections that are marked and cut, and render them one by one.

6. The method for quickly checking digital pathological sections and observing missed tissues in a pathological diagnosis system according to claim 1, characterized in that: The sliding track includes the following drawing steps: The server sets the initial cursor position coordinates and the moving position coordinates at each moment through the display image of the digital pathology slice; wherein, The display diagram is configured with a global coordinate system based on coordinate tracking; When the cursor coordinate position changes, the server generates a first response, and the first response information is used to indicate that the cursor has moved away from the previous coordinate position; After the first response is generated in the server, the cursor moves and generates a second response message, the second response message being used to indicate information that the cursor has entered any coordinate point of the digital pathology slice; The server generates a sliding track of the cursor according to the first response information and the second response message.

7. The method for quickly checking digital pathological sections and observing missed tissues in a pathological diagnosis system according to claim 6, characterized in that: The sliding track includes the following drawing steps: Receive the position point information of the initial position coordinates of the cursor on the digital pathology slice, and write the position point information into the cache queue; wherein, The location point information includes the timestamp when the user terminal draws the location point and the location information of the location point; A preset time length and a screening mechanism for each preset time length are set, and first target location point information corresponding to a timestamp matching the screening time is screened out from the location point information cached in the cache queue, and trajectory drawing is performed based on the first target location point information; If the preset duration of trajectory drawing is insufficient under the screening mechanism, trajectory drawing overlap is performed based on the location point information currently cached in the cache queue and the first target location point information drawn within the last preset duration.

8. The method for quickly checking digital pathological sections and observing missed tissues in a pathological diagnosis system according to claim 2, characterized in that: The screen stay includes: Acquire multiple frames of initial images of different regions of the digital pathology slice displayed on the screen; Perform region detection on each initial image to obtain time information of different regions on the screen; determining first intersection information and second intersection information of time information and area boundaries of different areas of the digital pathology slice, The first cross information is used to indicate the first time when the target area of ​​the digital pathology slice enters the screen display; The second intersection point is used to indicate a second time when the target area of ​​the digital pathology slice moves out of the screen display; The movement time of the target area between the first time and the second time is obtained, and the movement time is used as the residence time of the target area on the screen.

9. The method for quickly checking digital pathological sections and observing missed tissues in a pathological diagnosis system according to claim 1, characterized in that: The rendering further comprises: The server receives the browsing information for the target area sent by the application, determines whether the browsing is completed, and responds to the corresponding rendering command when the browsing is completed; According to the rendering instruction, the target area is rendered to obtain a rendering result of the target area; During the rendering process of the target area, the overall rendering image is drawn.

10. The method for quickly checking digital pathological sections and observing missed tissues in a pathological diagnosis system according to claim 9, characterized in that: The rendering command is generated based on a pathological matching mechanism; wherein, The pathology matching mechanism is connected with the pathology library, pathology feature library and rendering configuration matrix; The pathology matching mechanism matches information of different regions of digital pathology sections through different cancer types in the pathology library and pathology features in the pathology feature library; Information matching includes pathological feature matching and cancer type matching; According to the pathological feature matching results and the cancer type matching results, determining the rendering values ​​corresponding to the corresponding feature parameters and the cancer type parameters in the rendering configuration matrix; According to the rendering value, the rendering color and rendering depth are determined, and a rendering command is generated.

Images