Slice focal map acquisition method and device, computer device and storage medium

Abstract

The application discloses a slice focal point map acquisition method and device, computer equipment and a storage medium, and relates to the technical field of artificial intelligence. The method comprises the following steps: acquiring a slice scanning image of a target sampling point of the slice through a digital slice scanner; performing defocus information prediction based on the image content of the slice scanning image to obtain defocus information of the slice scanning image; the defocus information is used for indicating the deviation degree between the axial position of the target sampling point and the focal plane position of the target sampling point; calculating the focal plane position of the target sampling point according to the defocus information; and generating the focal point map based on the focal plane position of the target sampling point. The application can enable the digital slice scanner to quickly generate the focal point map of the slice in the pre-scanning process of the slice, and the accuracy of the focal point map is relatively high.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular to a method, apparatus, computer device and storage medium for obtaining a slice of focus map. Background Technology

[0002] Digital slicing scanners convert traditional physical slices into whole-slide images (WSI), making slice storage more convenient and allowing for subsequent analysis using artificial intelligence technologies. During the conversion process, to eliminate focusing errors caused by the slice's inherent undulations, the digital slicing scanner pre-scans the slice before the actual scan to obtain the focal positions of multiple sampling points, creating a focus map.

[0003] In related technologies, to obtain a focal map of a slice, each sampling point on the slice is progressively scanned axially. Specifically, the digital slice scanner sample platform is moved horizontally to a sampling position and then fixed. The axial position is gradually changed while scanning the sampling point multiple times. An image sharpness algorithm (such as the Brenner gradient algorithm, which calculates image sharpness by the square of the gray-level difference between two adjacent pixels) is used to determine the optimal focal plane. The axial position of the digital slice scanner sample platform corresponding to the acquisition of this optimal focal plane is recorded. By performing this axial progressive scanning on each sampling point on the slice, the focal map of the slice is determined.

[0004] However, because digital slicing scanners require axial step scanning and multiple slice scan images for each sampling point during the pre-scanning process, the pre-scanning time is long and the efficiency is low; but reducing the number of sampling points will reduce the accuracy of the focal map. How to obtain the focal map of the slice quickly while ensuring its accuracy is a technical problem that needs to be solved. Summary of the Invention

[0005] This application provides a method, apparatus, computer device, and storage medium for acquiring focus maps of slicing, enabling digital slicing scanners to quickly generate focus maps during pre-scanning while ensuring the accuracy of the focus map, thereby improving the speed of focus map generation. The technical solution is as follows:

[0006] According to one aspect of this application, a method for obtaining a focus map of a tile is provided, the method comprising:

[0007] A slice scan image of the target sampling points of the slice is obtained using a digital slice scanner;

[0008] Based on the image content of the sliced ​​scan image, defocus information is predicted to obtain the defocus information of the sliced ​​scan image; the defocus information is used to indicate the degree of deviation between the axial position of the target sampling point and the focal plane position of the target sampling point;

[0009] Calculate the focal plane position of the target sampling point based on the defocus information;

[0010] The focal map is generated based on the focal plane position of the target sampling point.

[0011] According to another aspect of this application, a slice-based focus map acquisition device is provided, the device comprising:

[0012] The acquisition module is used to acquire a slice scan image of the target sampling points of the slice using a digital slice scanner;

[0013] The prediction module is used to predict defocus information based on the image content of the sliced ​​scan image to obtain the defocus information of the sliced ​​scan image; the defocus information is used to indicate the degree of deviation between the axial position of the target sampling point and the focal plane position of the target sampling point.

[0014] The calculation module is used to calculate the focal plane position of the target sampling point based on the defocus information;

[0015] The map generation module is used to generate the focal map based on the focal plane position of the target sampling point.

[0016] According to another aspect of this application, a computer device is provided, the computer device including a processor and a memory, the memory storing at least one instruction, the instruction being loaded and executed by the processor to implement the slice focus map acquisition method as provided in various aspects of this application.

[0017] According to another aspect of this application, a computer-readable storage medium is provided, wherein computer instructions are stored therein, which are loaded and executed by a processor to implement the focus map acquisition method for slices as provided in various aspects of this application.

[0018] According to one aspect of this application, a computer program product is provided, comprising 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 slice focus map acquisition method.

[0019] The embodiments of this application include at least the following beneficial effects:

[0020] By acquiring a slice scan image of the target sampling point, defocus information is predicted from the image content of the slice scan image to obtain the defocus information of the target sampling point. Then, based on the defocus information, the focal plane position of the target sampling point is calculated to generate a focus map. This provides an efficient method for obtaining a focus map. Only one slice scan image needs to be acquired for each sampling point to predict the defocus information, saving the time spent on multiple axial scans for each sampling point. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments of this application 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.

[0022] Figure 1 This is a schematic diagram of a method for obtaining the focus map of a slice provided in an exemplary embodiment of this application;

[0023] Figure 2 This is a structural block diagram of a computer system provided in an exemplary embodiment of this application;

[0024] Figure 3 This is a flowchart of a training method for a defocus information prediction model provided in an exemplary embodiment of this application;

[0025] Figure 4 This is a schematic diagram of the training process of a defocus information prediction model provided in an exemplary embodiment of this application;

[0026] Figure 5 This is a flowchart of a method for obtaining the focus map of a tile, provided in an exemplary embodiment of this application;

[0027] Figure 6 This is a schematic diagram of a focus map provided in an exemplary embodiment of this application, and an effect diagram of scanning a slice based on the focus map;

[0028] Figure 7 This is a flowchart of a method for obtaining the focus map of a tile, provided in an exemplary embodiment of this application;

[0029] Figure 8 This is a flowchart of a method for obtaining the focus map of a tile, provided in an exemplary embodiment of this application;

[0030] Figure 9 This is a structural block diagram of a slice focus map acquisition device provided in an exemplary embodiment of this application;

[0031] Figure 10 This is a block diagram of a computer device provided in an exemplary embodiment of this application. Detailed Implementation

[0032] 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.

[0033] In the following description, when referring to the accompanying drawings, the same numbers in different drawings denote the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0034] In the description of this application, it should be understood that the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance. In the description of this application, it should be noted that, unless otherwise explicitly specified and limited, the terms "connected" and "linked" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances. Furthermore, in the description of this application, unless otherwise stated, "multiple" 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 existing alone, A and B existing simultaneously, or B existing alone. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship.

[0035] To facilitate understanding of the solutions shown in the embodiments of this application, the terms appearing in the embodiments of this application will be introduced below.

[0036] 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.

[0037] 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.

[0038] Machine Learning (ML) is a multidisciplinary field involving probability theory, statistics, approximation theory, convex analysis, and algorithm complexity theory. It specifically studies how computers can simulate or implement human learning behavior to acquire new knowledge or skills and reorganize existing knowledge structures to continuously improve their performance. Machine learning is the core of artificial intelligence and the fundamental way to endow computers with intelligence; its applications span all areas of artificial intelligence. Machine learning and deep learning typically include techniques such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and instruction-based learning.

[0039] Deep Learning (DL) is a new research direction in the field of machine learning. By learning the inherent patterns and representation levels of sample data, the information obtained during the learning process can greatly help in interpreting data such as text, images, and sound. Ultimately, it enables machines to have analytical and learning capabilities like humans, and to recognize data such as text, images, and sound.

[0040] Digital slides, also known as virtual slides, are obtained by scanning and converting physical slides, allowing the physical slides to be stored on computer devices and easily analyzed. Taking pathological slides as an example, digital slides contain all the lesion information from the physical slides.

[0041] 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.

[0042] The solutions provided in this application involve technologies such as neural networks in artificial intelligence, which will be specifically described in the following embodiments.

[0043] In related technologies, each sampling point of the slice is scanned multiple times by changing the axial position of the sample platform to obtain multiple slice scan images. The slice scan image with the highest image clarity is selected, and the axial position of the sample platform when acquiring this slice scan image is taken as the focal plane position of that sampling point. The focal plane position is the position of the sample platform that allows the digital slice scanner to accurately focus on the sampling point and obtain a clear slice scan image. Because multiple scans are performed on each sampling point, the pre-scan is very time-consuming; if the number of sampling points is reduced to reduce the pre-scan time, the accuracy of the focus map will decrease. Therefore, the focus map acquisition method in related technologies cannot achieve a balance between speed and high accuracy. The method proposed in this application obtains a slice scan image of the target sampling point and calls a defocus information prediction model to predict its defocus information, thus obtaining the focal plane position of the target sampling point. This allows for the rapid acquisition of a high-precision focus map during the pre-scanning process using a digital slice scanner.

[0044] Figure 1 A schematic diagram of a method for obtaining the focus map of a slice is shown. For the objective lens currently used in digital slicing scanners, a deep learning network is pre-trained to predict the defocus information of sampling points, i.e. Figure 1 The defocus information prediction network in 18.

[0045] like Figure 1 As shown, slide 14 is an extremely thin section suitable for microscopic examination or scanning with a digital slide scanner, such as a pathological slide, a material slide, etc. To obtain a focal map of the slide, several sampling points are pre-selected on slide 14, and these sampling points are scanned to acquire slide scan images. The plane where the sample platform 12 on which slide 14 is placed is taken as the XY horizontal plane, and the direction perpendicular to the XY horizontal plane is the axial direction, i.e., the Z direction. The positive direction of the Z direction is defined as the direction perpendicular to the XY horizontal plane and pointing towards the sky. By moving the sample platform 12 horizontally in the XY plane, different sampling points can be scanned; by moving the sample platform 12 in the Z direction, the same sampling point can be axially scanned, acquiring slide scan images of that sampling point at different axial positions.

[0046] For example, the sample platform 12 of the digital slice scanner is moved horizontally to the position of the objective lens 10 aligned with the i-th sampling point, and an image is acquired at the i-th sampling point to obtain the slice scan image 16. Then, the defocus information of the slice scan image 16 acquired at the i-th sampling point is calculated by calling the defocus information prediction network 18, and the defocus information of the i-th sampling point is obtained as: defocus amount Z = -3.5 μm.

[0047] For each sampling point on the slice, repeat the steps of acquiring the slice scan image and obtaining defocus information through the defocus information prediction network to obtain the defocus information of each sampling point. Finally, the discrete data points representing the defocus information of the sampling points in the 3D focus map are connected by interpolation calculation to obtain the focus map of the entire slice.

[0048] Figure 2 This illustration shows a schematic diagram of a computer device provided in an exemplary embodiment of this application. The computer device can be part of a digital slice scanner or a server connected to the digital slice scanner. The device includes: a bus 101, a processor 102, a memory 103, and a slice scanning image acquisition component 104.

[0049] The processor 102 includes one or more processing cores. The processor 102 executes various functional applications and information processing by running software programs and modules.

[0050] The memory 103 is connected to the processor 102 via the bus 101.

[0051] The memory 103 may be used to store at least one instruction, which the processor 102 may execute to implement the steps in the following method embodiments.

[0052] Furthermore, the memory 103 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, including but not limited to: magnetic disks or optical disks, electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), static random access memory (SRAM), read-only memory (ROM), magnetic storage, flash memory, and programmable read-only memory (PROM).

[0053] The slice scanning image acquisition component 104 can be used to acquire slice scanning images of target sampling points. For example, the slice scanning image acquisition component 104 can be the objective lens in a digital slice scanner.

[0054] Figure 3 A flowchart illustrating a training method for a defocus information prediction model provided in an exemplary embodiment of this application is shown. Exemplarily, this method can be... Figure 2 The computer device shown executes the method. The method includes the following steps:

[0055] Step 210: Obtain an axial scan image set using a digital slice scanner;

[0056] Axial scan image sets are acquired by axial step-scanning using a digital slice scanner.

[0057] For example, the scan step size is set to 0.5 μm. For any sampling point, the axial scan image set of the sampling point is obtained by scanning the sampling point axially every 0.5 μm using a digital slice scanner.

[0058] Optionally, the clearest image in the axial scan image set is determined by an image sharpness algorithm, and this image is used as the focused image. The axial position of the sample platform from which the focused image is obtained is used as the focal plane position with a defocus amount Z=0μm.

[0059] Step 220: Obtain the training sample set;

[0060] The training sample set includes: sample sampling point images and defocus information labels for the sample sampling points.

[0061] For example, for the axial scan image set obtained in step 210, a total of 21 scan images are taken, ranging from a defocus amount Z = -10μm to a defocus amount Z = +10μm, with a scan step size of 0.5μm. Each scan image has a defocus amount label corresponding to that scan image. Nine 672 images are randomly cut from each scan image. Image blocks of size 672, and assign the defocus amount label of the scanned image to each 672. A 672-pixel image block. This method yields a block containing 21... Nine sample point images, and a training sample set of defocus information labels for the sample point images corresponding to the sample point images.

[0062] For example, using the method described above to obtain training sample sets, n training sample sets are obtained. The larger the value of n, the higher the accuracy of the final trained model, but the greater the computational resources required. Illustratively, n is set to 1000, meaning 1000 training sample sets are obtained for training the defocus information prediction model.

[0063] Optionally, dataset augmentation operations can be performed to expand the dataset used for training, such as rotation and mirroring operations.

[0064] Step 230: Input the sampled point image into the defocus information prediction model to predict the defocus information and obtain the predicted defocus information of the sample;

[0065] Input the training sample set obtained in step 220 into the defocus information prediction model to obtain the predicted defocus information output by the defocus information prediction model.

[0066] The defocus information prediction model can be at least one of AlexNet, MobileNetV3, InceptionV3, and ResNet50.

[0067] Step 240: Calculate the error loss between the predicted out-of-focus information and the out-of-focus information label;

[0068] The error loss is obtained by subtracting the predicted defocus information obtained from the defocus information prediction model in step 230 from the defocus information labels of the sample sampling points in the training sample set.

[0069] Step 250: Train the defocus information prediction model based on error loss.

[0070] The model is trained based on the error loss between the predicted out-of-focus information and the out-of-focus information label obtained in step 240. That is, the predicted out-of-focus information obtained by the out-of-focus information prediction model continuously approaches the out-of-focus information label of the sample sampling point.

[0071] For example, the backpropagation algorithm is used to backpropagate the error loss of the out-of-focus information prediction model, and the model parameters in the out-of-focus information prediction model are updated. Training of the out-of-focus information prediction model ends when the training termination condition is met.

[0072] For example, training termination conditions include: the number of training iterations reaches a preset number, or the error loss decreases to below a preset threshold.

[0073] based on Figure 3 The flowchart shown illustrates the training method for the defocus information prediction model. Figure 4 A schematic diagram of the training process of a defocus information prediction model using the MobileNetV3 neural network as an example is shown.

[0074] The digital slice scanner 32 acquires an axial scan image set through axial step scanning. The axial scan image set undergoes operations such as cutting and dataset augmentation to obtain a training sample set 34, which includes sample sampling point images and defocus information labels for the sample sampling points. The training sample set 34 is input into the defocus information prediction model 36 to obtain predicted defocus information 38. The defocus information prediction model is trained by calculating the error loss between the predicted defocus information 38 and the defocus information labels in the training sample set 34.

[0075] Figure 5A flowchart illustrating a method for obtaining a focus map of a tile according to an exemplary embodiment of this application is shown. Exemplarily, this method may be... Figure 2 The computer device shown executes the method. The method includes the following steps:

[0076] Step 320: Obtain a slice scan image of the target sampling points of the slice using a digital slice scanner;

[0077] The target sampling points are sampling points pre-selected on the slice. Multiple sampling points can be randomly selected on the slice; or sampling points can be selected at fixed intervals; or, sampling points can be set in different ways in different areas of the slice according to the actual situation, etc. This application does not limit the number of sampling points or the method of selecting sampling points.

[0078] For example, the slide is placed on the sample platform of a digital slide scanner. For the i-th sampling point among the target sampling points, the slide scan image of the i-th sampling point is obtained by moving the sample platform of the digital slide scanner horizontally to the position where the objective lens is aligned with the i-th sampling point. Here, i is an integer greater than 1, and the i-th sampling point is any sampling point among the target sampling points.

[0079] Optionally, a reference focal plane position for the digital slicing scanner is pre-selected. The reference focal plane position can be a focal plane position already determined for any of the target sampling points on the slice; alternatively, the reference focal plane position can be a focal plane position calculated based on the previous sampling point. The sample platform of the digital slicing scanner is moved horizontally to the position where the objective lens is aligned with the i-th sampling point, and then axially moved to the reference focal plane position for scanning, thus acquiring a slice scan image of the i-th sampling point.

[0080] Step 340: Based on the image content of the sliced ​​scan image, predict the defocus information to obtain the defocus information of the sliced ​​scan image;

[0081] Defocus information is used to indicate the degree of deviation between the axial position of the target sampling point and the focal plane position of the target sampling point.

[0082] For example, a pre-trained defocus information prediction model is invoked to predict the defocus information of the image content of the slice scan image, thereby obtaining the defocus information of the slice scan image. The defocus information prediction model is a neural network model used to predict the defocus information.

[0083] Step 360: Calculate the focal plane position of the target sampling point based on the defocus information;

[0084] The defocus information obtained in the above steps is acquired by scanning the target sampling point using a digital slice scanner to obtain a slice scan image. Therefore, the defocus information obtained from this slice scan image is relative to the axial position of the sample platform when the slice scan image was obtained. To obtain the focal plane position of the target sampling point, i.e., the absolute axial position of the focal plane of the target sampling point, it is necessary to calculate it by combining the reference focal plane position and the defocus information.

[0085] For example, given the focal plane position of a reference point, the Z-axis coordinate of the sample platform of the digital slice scanner at the focal plane position of the reference point is recorded as a μm, where a μm is the absolute axial position of the focal plane of the reference point. The axial position of the sample platform of the digital slice scanner is fixed at a μm and remains unchanged; this position is defined as Z=0 μm. When the defocus information of the i-th sampling point is obtained from the slice scan image of the i-th sampling point as a defocus amount Z=b μm, the focal plane position of the i-th sampling point can be calculated as (a+b) μm based on the defocus information and the focal plane position of the reference point, where b is a positive or negative value, and the sign of b indicates whether the focal plane position of the i-th sampling point is above or below the absolute axial position of the focal plane of the reference point.

[0086] For example, the focal plane position calculated at the i-th sampling point is determined as the reference focal plane position of the (i+1)-th sampling point. Given that the focal plane position of the first sampling point is a μm, it is used as the reference focal plane position of the second sampling point, i.e., the focal plane position of the first sampling point is defined as Z=0 μm. The defocus information of the second sampling point is obtained as the defocus amount Z=b μm through the slice scan image of the second sampling point. Based on the defocus information and the reference focal plane position, the focal plane position of the second sampling point can be calculated as (a+b) μm. Next, the focal plane position of the second sampling point is used as the reference focal plane position of the third sampling point, i.e., the focal plane position of the second sampling point is defined as Z=0 μm. The defocus information of the third sampling point is obtained as the defocus amount Z=c μm through the slice scan image of the third sampling point. Based on the defocus information and the reference focal plane position, the focal plane position of the third sampling point can be calculated as (a+b+c) μm. Subsequent sampling points follow the same principle, calculating the focal plane position of the target sampling point based on the defocus information and the reference focal plane position. Where b and c are both positive or negative values. The sign of b indicates that the focal plane position of the second sampling point is above or below the focal plane position of the first sampling point, and the sign of c indicates that the focal plane position of the third sampling point is above or below the focal plane position of the second sampling point, and so on for subsequent defocusing amounts.

[0087] Step 380: Generate a focal map based on the focal plane position of the target sampling point.

[0088] The above steps yielded the focal plane position of the target sampling point, which is the absolute position of the sample platform axis of the digital slice scanner. Since the focal plane position of the sampling point is a discrete data point, interpolation is needed to generate a continuous focal map in space. Interpolation involves interpolating a continuous function onto the discrete data, ensuring that the continuous surface or curve passes through all given discrete data points. Interpolation methods can include interpolation using two neighboring points, linear interpolation, cubic spline interpolation, etc. This application does not limit the interpolation method used.

[0089] A continuous focal map is generated by interpolating the focal plane positions of discrete target sampling points, such as... Figure 6 The focal map is shown in 22. Figure 6 The slice scan image 24 was obtained during the pre-scan. By calling the defocus information prediction model, the defocus amount of this slice scan image was calculated to be Z = 3.8 μm. Combined with the reference focal plane position, the focal plane position of this sampling point was obtained and recorded in the focus map. During the actual scan, the digital slice scanner controls the axial movement of the sample platform according to the defocus amount indicated in the focus map, acquiring the final slice scan image 26. Figure 6 A comparison of slice scan image 24 obtained during the pre-scanning process and slice scan image 26 obtained after axial position correction by focus map in the formal scan shows that after focus map correction, the digital slice scanner can obtain clearer slice scan images in the formal scan.

[0090] In summary, the method provided in this embodiment predicts the defocus information of the slice scan image of the target sampling point of the slice obtained by the digital slice scanner, obtains the defocus information of the slice scan image, and then obtains the focal plane position of the target sampling point based on the defocus information. Based on the focal plane position, a focus map is generated, which provides a method for obtaining the focus map of the slice that saves pre-scanning time and is convenient and efficient.

[0091] The method provided in this application can significantly reduce the time spent in the pre-scanning process of a digital slice scanner, thereby greatly improving the scanning efficiency of the digital slice scanner.

[0092] For example, assume the exposure time t for acquiring each slice of scanned image. 曝光 = 20 milliseconds, unit step t of mechanical axial movement z_move and mechanical horizontal movement t xy_move The time taken is 100 milliseconds. Related techniques require acquiring approximately 15 slice scan images for each sampling point, therefore the time required for each sampling point is t1 = (t... z_move +t 曝光 ) 15 = (100 + 20) 15 = 1800 ms. Taking 25 sampling points from a slice as an example, the total time required is t. 1_total =(t1+t xy_move ) 25 = 47500 ms.

[0093] The method provided in this application only requires axial scanning at the reference point, acquiring 15 slice scan images, and for the remaining sampling points, only one slice scan image is needed to predict defocus information. That is, for the reference point, the time t... 2_refer = t1=1800 ms; for the other reference points, the time taken is only t2=t 曝光 1 = 20 ms. Similarly, taking 25 sampling points from a slice as an example, the total time required is t. 2_total = t 2_refer + t2 24 + t xy_move 25 = 4780 ms.

[0094] From the above calculations, we can obtain t 1_total / t 2_total The speed of acquiring the focus map during the pre-scanning process using the method provided in this application is approximately 10 times faster than that of acquiring the focus map in related technologies. Furthermore, the more sampling points are taken for a slice, the more pronounced the advantage of the method provided in this application becomes.

[0095] Before scanning the target sampling points on the slice to obtain the slice scan image, a reference point is first selected, and the focal plane position of the reference point is obtained. This focal plane position is then used as the reference focal plane position, and subsequent scanning of the target sampling points is performed to obtain the defocus information of the target sampling points based on the reference focal plane position. This process is described in detail in the following embodiments.

[0096] Figure 7 A flowchart illustrating a method for obtaining a focus map of a tile according to an exemplary embodiment of this application is shown. Exemplarily, this method may be... Figure 2 The computer device shown executes the method. The method includes the following steps:

[0097] Step 410: Perform an axial scan on the reference point to obtain the focal plane position of the reference point;

[0098] The reference point can be any one of the target sampling points; or, the reference point can be the first sampling point scanned by the digital slice scanner in a certain order, and so on. This application does not restrict the method of selecting the reference point.

[0099] A set of slice scan images of a reference point along the axial direction is acquired using a digital slice scanner. This set of slice scan images includes at least two slice scan images acquired at fixed scanning intervals. The target slice scan image with the highest resolution in the slice scan image set is determined. The acquisition position of the target slice scan image is determined as the focal plane position of the reference point.

[0100] For example, an axial step scan is performed on the reference point to obtain the focal plane position of the reference point.

[0101] For example, the first sampling point scanned by a digital slice scanner in a certain sequence is selected as the reference point. Scanning is performed at a fixed interval of 0.5 μm to obtain a set of slice scan images along the axis of the reference point. An image sharpness algorithm, such as the Brenner gradient image sharpness algorithm, is used to calculate the image sharpness of the slice scan image set, determining the slice scan image with the highest image sharpness. The axial position of the sample platform when acquiring this slice scan image is obtained, and this axial position is determined as the focal plane position of the reference point.

[0102] Step 422: Adjust the axial position of the sample platform of the digital slice scanner to the reference focal plane position;

[0103] The focal plane position of the reference point is obtained through the above steps. This focal plane position is used as the reference focal plane position for the digital slicing scanner to scan subsequent sampling points. The sample platform of the digital slicing scanner is then moved axially to the reference focal plane position.

[0104] Step 424: Obtain the slice scan image of the i-th sampling point of the sample platform at the reference focal plane position using a digital slice scanner;

[0105] In step 422, the sample platform has been adjusted to the reference focal plane position. In this step, the sample platform is moved horizontally until the objective lens is aligned with the i-th sampling point, and the i-th sampling point is scanned using a digital slice scanner to obtain a slice scan image of the i-th sampling point.

[0106] Step 442: Call the defocus information prediction model to predict the defocus information of the slice scan image at the i-th sampling point, and obtain the defocus information of the slice scan image;

[0107] The defocus information prediction model is a pre-trained neural network model designed for the magnification of the objective lens currently used in the digital slide scanner. The specific steps for training the defocus information prediction model can be found in step 340, and will not be repeated here.

[0108] The defocus information prediction model is invoked to predict the defocus information of the slice scan image at the i-th sampling point obtained in step 424. The defocus information prediction model predicts the defocus information based on the image content of the slice scan image at the i-th sampling point, and obtains the defocus information of the slice scan image, for example, the defocus amount at the i-th sampling point.

[0109] After performing defocus information prediction on the slice scan image of the i-th sampling point in step 442, perform the same operation on the (i+1)-th sampling point, that is, make i=i+1, and repeat step 424 until all sampling points have completed defocus information prediction.

[0110] Step 462: Using the reference focal plane position as a reference and the defocus information predicted by the slice scan image as the relative defocus information, calculate the focal plane position of the i-th sampling point;

[0111] The defocus information obtained in step 442 is calculated relative to the reference focal plane position. Therefore, in order to obtain the focal plane position of the i-th sampling point, that is, the absolute axial position of the focal plane of the i-th sampling point, it is necessary to add the reference focal plane position and the defocus information.

[0112] For example, when the focal plane position of the reference point is obtained in step 410, the absolute axial position of the focal plane position of the reference point is recorded as a μm, and the focal plane position of the reference point is defined as the reference focal plane position, that is, the defocus amount of the plane is Z=0 μm. When the defocus information of the i-th reference point is obtained as defocus amount Z=b μm in step 442, the focal plane position of the i-th reference point is calculated as (a+b) μm.

[0113] Step 480: Generate a focal map based on the focal plane position of the target sampling point.

[0114] The above steps yielded the focal plane position of the target sampling point, which is the absolute position of the sample platform axis of the digital slice scanner. Since the focal plane position of the sampling point is discrete, interpolation is needed to generate a continuous focal map in space. Interpolation involves interpolating a continuous function onto discrete data, ensuring that a continuous surface or curve passes through all given discrete data points. Interpolation methods can include interpolation using two neighboring points, linear interpolation, cubic spline interpolation, etc. This application does not limit the interpolation method used.

[0115] In summary, the focus map acquisition method for slices provided in this embodiment first obtains the focal plane position of the reference point by axially scanning the reference point, then fixes the sample platform axially at the focal plane position, and then scans and predicts subsequent sampling points to generate a focus map. This provides a method for acquiring focus maps of slices without axially moving the sample platform during the entire pre-scanning process, enabling the pre-scanning process to be realized quickly and improving the efficiency of digital slice scanners in obtaining focus maps of slices.

[0116] For slices with large surface undulations, a method based on... Figure 7 The method described in the embodiment for acquiring the focus map may encounter problems, such as the axial position difference between the sampling point and the reference point exceeding the range that the defocus information prediction network can accurately predict. This problem can be addressed by adjusting the position of the reference focal plane during the pre-scanning process.

[0117] Figure 8 A flowchart illustrating a method for obtaining a focus map of a tile according to an exemplary embodiment of this application is shown. Exemplarily, this method may be... Figure 2 The computer device shown executes the method. The method includes the following steps:

[0118] Step 510: Perform an axial scan on the first reference point to obtain the focal plane position of the first reference point;

[0119] The first reference point can be any one of the target sampling points; or, the first reference point can be the first sampling point scanned by the digital slice scanner in a certain order, and so on. This application does not limit the method of selecting the first reference point.

[0120] A set of slice scan images of a first reference point along the axial direction is acquired using a digital slice scanner. The set of slice scan images includes at least two slice scan images acquired at fixed scanning intervals. The target slice scan image with the highest resolution in the slice scan image set is determined. The acquisition position of the target slice scan image is determined as the focal plane position of the first reference point.

[0121] For example, the first sampling point scanned by a digital slice scanner in a certain sequence is selected as the first reference point. Scanning is performed at a fixed interval of 0.5 μm to obtain a set of slice scan images along the axis of the first reference point. The Brenner gradient image sharpness algorithm is used to calculate the image sharpness of the slice scan image set, and the slice scan image with the highest image sharpness is determined. The acquisition position of the sample platform when acquiring this slice scan image is obtained, and this position is determined as the focal plane position of the first reference point.

[0122] Step 522: Adjust the axial position of the sample platform of the digital slice scanner to the focal plane position of the first reference point;

[0123] The focal plane position of the first reference point is obtained through the above steps. This focal plane position is used as the reference focal plane position for the digital slicing scanner to scan subsequent sampling points. The sample platform of the digital slicing scanner is moved axially to the focal plane position of the first reference point.

[0124] Step 524: Obtain the slice scan image of the sample platform at the reference focal plane position corresponding to the i-th sampling point using a digital slice scanner;

[0125] In step 522, the sample platform has been adjusted to the reference focal plane position, i.e., the focal plane position of the first reference point. In this step, the sample platform is moved horizontally until the objective lens is aligned with the i-th sampling point, and the i-th sampling point is scanned using a digital slice scanner to obtain a slice scan image of the i-th sampling point.

[0126] Step 542: Call the defocus information prediction model to predict the defocus information of the slice scan image at the i-th sampling point, and obtain the defocus information of the slice scan image;

[0127] The defocus information prediction model is a pre-trained neural network model for the magnification of the currently used objective lens. The specific steps for training the defocus information prediction model can be found in step 340, and will not be repeated here.

[0128] The defocus information prediction model is invoked to predict the defocus information of the slice scan image at the i-th sampling point obtained in step 424. The defocus information prediction model predicts the defocus information based on the image content of the slice scan image at the i-th sampling point, and obtains the defocus information of the slice scan image, for example, the defocus amount at the i-th sampling point.

[0129] Step 544: Determine whether the absolute value of the defocus information predicted by the slice scan image at the i-th sampling point is greater than the absolute value of a predetermined threshold;

[0130] The defocus information prediction model used in the above steps is trained based on the sample sampling point image and the defocus information label of the sample sampling point image. The defocus information label of the sample sampling point image belongs to the target value range, and the predetermined threshold is determined based on the target value range.

[0131] For example, the training sample set used when training the defocus information prediction model includes sampled point images of the defocus amount in the range of Z=-10 μm to Z=10 μm and the defocus information label corresponding to the sampled point image. The defocus information label belongs to the target value range, that is, ±10 μm.

[0132] For example, a predetermined threshold is determined to be a value not exceeding half of the target value range, such as ±2 μm, ±3 μm, ±5 μm, etc. The smaller the absolute value of the selected predetermined threshold, the higher the accuracy of the focal map of the final slice; however, this leads to an increase in processing time due to more frequent axial position movements of the sample platform.

[0133] In step 542, the defocus information prediction model is invoked to predict the defocus information of the i-th sampling point. The absolute value of the defocus information of the i-th sampling point is compared with the absolute value of a predetermined threshold. If the absolute value of the defocus information of the i-th sampling point is greater than the absolute value of the predetermined threshold, step 546 is executed; if the absolute value of the defocus information of the i-th sampling point is not greater than the absolute value of the predetermined threshold, the axial position of the sample platform remains unchanged, and defocus information prediction is performed for the (i+1)-th sampling point, i.e., i = i+1, and step 524 is executed again until prediction is completed for all sampling points.

[0134] Step 546: Update the focal plane position of the i-th sampling point to the reference focal plane position corresponding to the (i+1)-th sampling point;

[0135] If the absolute value of the defocus information at the i-th sampling point is greater than the absolute value of a predetermined threshold, this step is executed.

[0136] Step 542 yields the defocus information for the i-th sampling point. An absolute value greater than a predetermined threshold indicates that the slice surface may have significant undulations near the i-th sampling point. To ensure that predictions for subsequent sampling points do not exceed the prediction range of the defocus information prediction model (i.e., the target value range to which the defocus information labels in the training sample set used to train the model belong), the focal plane position of the i-th sampling point is updated to the reference focal plane position corresponding to the (i+1)-th sampling point.

[0137] For example, a predetermined threshold is set to ±3 μm. The absolute position of the reference focal plane when acquiring the slice image of the i-th sampling point is a μm. The defocus amount of the i-th sampling point is predicted as Z=b μm by the defocus information prediction model. If the absolute value of b is greater than 3, the focal plane position of the i-th sampling point is updated to the reference focal plane position corresponding to the (i+1)-th sampling point. That is, the sample platform of the digital slice scanner is axially moved to the focal plane position of the i-th sampling point, the absolute position of the focal plane position of the i-th sampling point is recorded as (a+b) μm, and the focal plane position of the i-th sampling point is defined as the reference focal plane position with a defocus amount of Z=0 μm. Sampling prediction operation is then performed on the (i+1)-th sampling point.

[0138] After completing step 546, predict the out-of-focus information for the (i+1)th sampling point, i.e., make i = i+1, and repeat step 524 until prediction is completed for all sampling points.

[0139] Step 562: Using the reference focal plane position as a reference and the defocus information predicted by the slice scan image as the relative defocus information, calculate the focal plane position of the i-th sampling point;

[0140] The defocus information obtained in step 542 is the defocus information calculated relative to the reference focal plane position. Therefore, in order to obtain the focal plane position of the i-th sampling point, that is, the absolute axial position of the focal plane of the i-th sampling point, it is necessary to combine the reference focal plane position and the defocus information for addition calculation.

[0141] For example, when obtaining the focal plane position of the reference point in step 510, the absolute axial position of the focal plane position of the reference point is recorded as a μm, and the focal plane position of the reference point is defined as the reference focal plane position, that is, the defocus amount of the plane is Z=0 μm. In step 542, when the defocus information of the i-th sampling point is obtained as the defocus amount Z=b μm, the focal plane position of the i-th sampling point is calculated as (a+b) μm. When the absolute value of b is greater than the absolute value of a predetermined threshold, the focal plane position (a+b) μm of the i-th sampling point is taken as the reference focal point position of the (i+1)-th sampling point, and the defocus information of the (i+1)-th sampling point is obtained as the defocus amount Z=c μm, and the focal plane position of the (i+1)-th sampling point is obtained as (a+b+c) μm.

[0142] Step 580: Generate a focal map based on the focal plane position of the target sampling point.

[0143] The above steps yielded the focal plane position of the target sampling point, which is the absolute position of the sample platform axis of the digital slice scanner. Since the focal plane position of the sampling point is discrete, interpolation is needed to generate a continuous focal map in space. Interpolation involves interpolating a continuous function onto discrete data, ensuring that a continuous surface or curve passes through all given discrete data points. Interpolation methods can include interpolation using two neighboring points, linear interpolation, cubic spline interpolation, etc. This application does not limit the interpolation method used.

[0144] Optionally, step 544 in this embodiment can be omitted; that is, after step 542 is executed, steps 546, 562, and 580 are executed sequentially. In this execution method, after predicting the defocus information for the i-th sampling point, the sample platform is moved to the predicted focal plane position of the i-th sampling point, which serves as the reference focal plane position for the (i+1)-th sampling point. The slice scan image of the (i+1)-th sampling point is then acquired at this reference focal plane position. The focus map obtained using this method has high accuracy, but because the sample platform of the digital slice scanner may need to be moved axially before predicting the defocus information for each sampling point, the process is time-consuming.

[0145] In summary, the slice focus map acquisition method provided in this embodiment updates the reference focal plane position when the axial positions of two sampling points differ significantly. Based on the updated reference focal plane position, it acquires the slice scan image of subsequent sampling points and calculates the defocus information of the slice scan image. This provides a method for acquiring a slice focus map by adjusting the axial position of the sample platform according to actual conditions. For slices with large surface undulations, updating the reference focal plane position ensures that the axial position of the sampling points does not exceed the prediction range of the defocus prediction model, resulting in higher accuracy of the acquired slice focus map.

[0146] In hospitals, doctors obtain information about patients' conditions by analyzing pathological slides. This requires the use of digital slide scanners to convert physical pathological slides into digital slides, facilitating analysis by doctors using computer equipment. The digital slide scanner first performs a pre-scan of the pathological slide, obtaining a focal map during this process. Then, it performs a formal scan, using the defocus information indicated by the focal map to move the axial position of the sample platform to the focal plane, acquiring a clear slide image. This image is then converted into a digital slide and stored in the computer.

[0147] When pre-scanning pathological slides using the focus map acquisition method of this application, two approaches can be used to quickly obtain high-precision focus maps. For pathological slides with minimal surface undulations, axial scanning of reference points on the slide is performed, and the focal plane position of the reference points is used as the reference focal plane position. Based on the reference focal plane position, defocus information prediction for all subsequent points is completed. This method can obtain the focus map of the pathological slide in the shortest possible time. For pathological slides with significant surface undulations, the reference focal plane position is changed according to the actual situation. When the axial position difference between adjacent sampling points exceeds a predetermined value, the reference focal plane position is updated to ensure that the prediction of defocus information for subsequent sampling points is within the range that the defocus information prediction model can accurately predict. This ensures high accuracy of the focus map while rapidly generating the focus map of the pathological slide.

[0148] 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.

[0149] Figure 9 This is a structural block diagram of a slice focus map acquisition device provided in an exemplary embodiment of this application. The device includes:

[0150] The acquisition module 920 is used to acquire a slice scan image of the target sampling points of the slice using a digital slice scanner;

[0151] The prediction module 940 is used to predict defocus information based on the image content of the sliced ​​scan image to obtain the defocus information of the sliced ​​scan image; the defocus information is used to indicate the degree of deviation between the axial position of the target sampling point and the focal plane position of the target sampling point.

[0152] Calculation module 960 is used to calculate the focal plane position of the target sampling point based on the defocus information;

[0153] The map generation module 980 is used to generate the focal map based on the focal plane position of the target sampling point.

[0154] In one possible design, the prediction module 940 is used to call the defocus information prediction model to predict the defocus information of the image content of the slice scan image, so as to obtain the defocus information of the slice scan image; wherein, the defocus information prediction model is a neural network model for predicting defocus information.

[0155] In one possible design, the defocus information prediction model is trained using the following steps: acquiring an axial scan image set using the digital slice scanner; acquiring a training sample set, which includes: sample sampling point images and defocus information labels for the sample sampling points; inputting the sample sampling point images into the defocus information prediction model to predict defocus information, thereby obtaining the predicted defocus information for the samples; calculating the error loss between the predicted defocus information and the defocus information labels; and training the defocus information prediction model based on the error loss.

[0156] In one possible design, the acquisition module 920 is used to adjust the axial position of the sample platform of the digital slice scanner to a reference focal plane position, which is a focal plane position already determined for a sampling point in the slice; and to acquire a slice scan image of the i-th sampling point of the sample platform at the reference focal plane position using the digital slice scanner; the calculation module 960 is used to calculate the focal plane position of the i-th sampling point based on the reference focal plane position and using the defocus information predicted by the slice scan image as relative defocus information.

[0157] In one possible design, the reference focal plane position includes the focal plane position of the first reference point in the slice; the acquisition module 920 is further configured to acquire a slice scan image set of the first reference point in the axial direction through the digital slice scanner, the slice scan image set including at least two slice scan images acquired at a fixed scanning interval; determine the target slice scan image with the highest clarity in the slice scan image set; and determine the acquisition position of the target slice scan image as the focal plane position of the first reference point.

[0158] In one possible design, the acquisition module 920 is further configured to update the focal plane position of the i-th sampling point to the reference focal plane position corresponding to the (i+1)-th sampling point.

[0159] In one possible design, the defocus information prediction is performed by a defocus information prediction model, which is trained based on sampled point images and defocus information labels of the sampled point images, wherein the defocus information labels belong to a target value range, and the predetermined threshold is determined based on the target value range.

[0160] In one possible design, the defocus information prediction model can be at least one of AlexNet, MobileNetV3, InceptionV3, and ResNet50.

[0161] Figure 10 This is a schematic diagram of a terminal according to an exemplary embodiment. The terminal 1000 includes a Central Processing Unit (CPU) 1001, a system memory 1004 including Random Access Memory (RAM) 1002 and Read-Only Memory (ROM) 1003, and a system bus 1005 connecting the system memory 1004 and the CPU 1001. The computer device 1000 also includes a basic input / output system (I / O system) 1006 to facilitate information transfer between various devices within the computer device, and a mass storage device 1007 for storing the operating system 1013, application programs 1014, and other program modules 1015.

[0162] The basic input / output system 1006 includes a display 1008 for displaying information and an input device 1009 for user input, such as a mouse or keyboard. Both the display 1008 and the input device 1009 are connected to the central processing unit 1001 via an input / output controller 1010 connected to the system bus 1005. The basic input / output system 1006 may also include the input / output controller 1010 for receiving and processing input from multiple other devices such as a keyboard, mouse, or electronic stylus. Similarly, the input / output controller 1010 also provides output to a display screen, printer, or other types of output devices.

[0163] The mass storage device 1007 is connected to the central processing unit 1001 via a mass storage controller (not shown) connected to the system bus 1005. The mass storage device 1007 and its associated computer device-readable media provide non-volatile storage for the computer device 1000. That is, the mass storage device 1007 may include computer device-readable media (not shown), such as a hard disk or a compact disc read-only memory (CD-ROM) drive.

[0164] Without loss of generality, the computer device readable medium may include computer device storage media and communication media. Computer device storage media include volatile and non-volatile, removable and non-removable media implemented using any method or technology for storing information such as computer device readable instructions, data structures, program modules, or other data. Computer device storage media include RAM, ROM, erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), CD-ROM, digital video disc (DVD), 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 device storage media are not limited to the above-mentioned types. The system memory 1004 and mass storage device 1007 described above can be collectively referred to as memory.

[0165] According to various embodiments of this disclosure, the computer device 1000 can also be connected to a remote computer device on a network, such as the Internet. That is, the computer device 1000 can be connected to the network 1011 via a network interface unit 1012 connected to the system bus 1005, or the network interface unit 1012 can be used to connect to other types of networks or remote computer device systems (not shown).

[0166] The memory also includes one or more programs stored in the memory, and the central processing unit 1001 executes the one or more programs to implement all or part of the steps of the above-mentioned slice focus map acquisition method.

[0167] In an exemplary embodiment, a computer-readable storage medium is also provided, which stores at least one instruction, at least one program, code set, or instruction set, 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 slice focus map acquisition method provided in the above-described method embodiments.

[0168] In an exemplary embodiment, a computer program product or computer program is also provided, which includes computer instructions stored in a computer-readable storage medium. A processor of a communication device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the communication device to perform the slice focus map acquisition method described above.

[0169] Those skilled in the art will understand that all or part of the steps of the above embodiments can be implemented by hardware or by a program instructing related hardware. The program can be stored in a computer-readable storage medium, such as a read-only memory, a disk, or an optical disk.

[0170] The above description is merely an optional 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 obtaining a sliced ​​focus map, characterized in that, The slice includes target sampling points, which are sampling points pre-selected on the slice. Each target sampling point includes the i-th sampling point in the slice, where the i-th sampling point is any one of the target sampling points, and i is an integer greater than 1. The method includes: The axial position of the sample platform of the digital slice scanner is adjusted to the reference focal plane position, and the slice scan image of the i-th sampling point of the sample platform at the reference focal plane position is obtained by the digital slice scanner. The reference focal plane position is the focal plane position that has been determined for a sampling point in the slice. A pre-trained defocus information prediction model is invoked to predict the defocus information of the slice scan image content, thereby obtaining the defocus information of the slice scan image; the defocus information is used to indicate the degree of deviation between the axial position of the target sampling point and the focal plane position of the target sampling point; Using the reference focal plane position as a reference and the defocus information predicted by the slice scan image as the relative defocus information, the focal plane position of the i-th sampling point is calculated; The focal map is generated based on the focal plane position of the target sampling point.

2. The method according to claim 1, characterized in that, The defocus information prediction model is trained using the following steps: Axial scan image set is obtained using the digital slice scanner; Obtain a training sample set, which includes: sample sampling point images and defocus information labels of the sample sampling points; The sampled image is input into the defocus information prediction model to predict the defocus information and obtain the predicted defocus information of the sample. Calculate the error loss between the predicted defocus information and the defocus information label; The defocus information prediction model is trained based on the error loss.

3. The method according to claim 1, characterized in that, The reference focal plane position includes the focal plane position of the first reference point in the slice; The method further includes: The digital slice scanner acquires a slice scan image set of the first reference point along the axial direction, the slice scan image set including at least two slice scan images acquired at a fixed scanning interval; Identify the target slice scan image with the highest resolution within the slice scan image set; The acquisition position of the target slice scan image is determined as the focal plane position of the first reference point.

4. The method according to claim 3, characterized in that, The first reference point is any sampling point other than the i-th sampling point among the target sampling points, or the first reference point is the first sampling point scanned by the digital slice scanner in a certain order.

5. The method according to claim 1, characterized in that, The method further includes: The focal plane position of the i-th sampling point is updated to the reference focal plane position corresponding to the (i+1)-th sampling point.

6. The method according to claim 5, characterized in that, The method further includes: If the defocus information predicted by the slice scan image is greater than a predetermined threshold, the focal plane position of the i-th sampling point is updated to the reference focal plane position corresponding to the (i+1)-th sampling point.

7. The method according to claim 6, characterized in that, The defocus information prediction is performed by a defocus information prediction model, which is trained based on sampled point images and the defocus information labels of the sampled point images. The defocus information labels of the sampled point images belong to the target value range, and the predetermined threshold is determined based on the target value range.

8. The method according to any one of claims 1 to 7, characterized in that, The defocus information prediction model is at least one of AlexNet, MobileNetV3, InceptionV3, and ResNet50.

9. A device for acquiring a sliced ​​focus map, characterized in that, The slice includes target sampling points, which are sampling points pre-selected on the slice. Each target sampling point includes the i-th sampling point in the slice, where the i-th sampling point is any one of the target sampling points, and i is an integer greater than 1. The device includes: The acquisition module is used to adjust the axial position of the digital slice scanner to the reference focal plane position, and to acquire the slice scan image of the i-th sampling point by means of the digital slice scanner at the reference focal plane position, wherein the reference focal plane position is the focal plane position that has been determined for a sampling point in the slice; The prediction module is used to call a pre-trained defocus information prediction model to predict the defocus information of the image content of the sliced ​​scan image, and obtain the defocus information of the sliced ​​scan image; the defocus information is used to indicate the degree of deviation between the axial position of the target sampling point and the focal plane position of the target sampling point; The calculation module is used to calculate the focal plane position of the i-th sampling point based on the reference focal plane position and the defocus information predicted by the slice scan image as the relative defocus information. The map generation module is used to generate the focal map based on the focal plane position of the target sampling point.

10. The apparatus according to claim 9, characterized in that, The defocus information prediction model is trained using the following steps: Axial scan image set is obtained using the digital slice scanner; Obtain a training sample set, which includes: sample sampling point images and defocus information labels of the sample sampling points; The sampled image is input into the defocus information prediction model to predict the defocus information and obtain the predicted defocus information of the sample. Calculate the error loss between the predicted defocus information and the defocus information label; The defocus information prediction model is trained based on the error loss.

11. A computer device, characterized in that, The device includes a processor, a memory connected to the processor, and program instructions stored in the memory, wherein the program instructions executed by the processor implement the slice focus map acquisition method as described in any one of claims 1 to 8.

12. A computer-readable storage medium storing program instructions, characterized in that, When the program instructions are executed by the processor, they implement the focus map acquisition method for slices as described in any one of claims 1 to 8.

13. A computer program product comprising computer instructions stored in a computer-readable storage medium, wherein a processor of a computer device reads the computer instructions from the computer-readable storage medium, and wherein the computer instructions, when executed by the processor, implement the slice focus map acquisition method as described in any one of claims 1 to 8.

Images