Tissue block measuring method and system based on digital image processing and computer vision

Abstract

The invention discloses a tissue block measurement method and system based on digital image processing and computer vision, and belongs to the field of measurement, and the method comprises the following steps: data acquisition: obtaining image data of a tissue block on a shooting platform; image preprocessing: preprocessing the collected image data, including denoising, distortion correction and white balance; image segmentation and measurement: identifying the contour of a tissue block in the processed image data, and measuring the length, width, area and perimeter of the tissue block; the measured data is managed, and a calculation result is associated with patient information and stored in a hospital information system; and displaying a calculation result in real time or generating a report and outputting to a specified terminal. According to the tissue block measurement method and system based on digital image processing and computer vision provided by the invention, the correction precision can be improved, the system adaptability is enhanced, and the measurement error is reduced.

Description

Technical Field

[0001] The present invention relates to the field of measurement technology, and in particular to a tissue block measurement method and system based on digital image processing and computer vision. Background Art

[0002] Digital image processing refers to the technology of analyzing, transforming and enhancing images through computer algorithms after digitizing them. It is widely used in many fields such as medical imaging, remote sensing, industrial detection, etc. In tissue block measurement, digital image processing technology is used to pre-process tissue sample images taken under a microscope, including noise removal, contrast enhancement, edge detection, etc., to provide a high-quality image foundation for subsequent feature extraction and measurement.

[0003] Computer Vision (CV) is a technology that enables computers to understand and interpret the content of images or videos. It uses machine learning, deep learning and other algorithms to extract useful information from images, such as shape, texture, color and other features, and classify, identify and measure them. In tissue block measurement, computer vision technology is used to automatically identify and segment tissue structures, quantify parameters such as cell number, size, distribution, etc., to achieve efficient and accurate measurement.

[0004] In the fields of biomedical research, pathological diagnosis, drug development, etc., accurate measurement of the size, morphology and distribution of tissue blocks is crucial for understanding disease mechanisms and evaluating treatment effects. Traditional manual measurement methods are not only time-consuming and laborious, but also easily affected by subjective factors, resulting in inconsistent results. Therefore, automated, high-precision tissue block measurement methods have become an urgent need. Summary of the invention

[0005] The purpose of the present invention is to provide a tissue block measurement method and system based on digital image processing and computer vision to solve the problems existing in the background technology.

[0006] To achieve the above object, the present invention provides a tissue block measurement system based on digital image processing and computer vision, comprising: A data acquisition module, including a shooting platform, is used to obtain image data of the tissue block on the shooting platform; Image preprocessing module, used for performing denoising, distortion correction, and white balance preprocessing operations on image data; Image segmentation and measurement module, used to identify the outline of tissue blocks and calculate the length, width, area and perimeter of tissue blocks; A data management module is used to associate the calculation results with the patient information and store them in the hospital information system; The result output module is used to display the calculation results in real time or generate a report and output it to the specified terminal.

[0007] Preferably, the result output module includes: Visual interface, used to display measured values ​​such as length, width, area and perimeter; Automatically generate measurement reports in PDF or other formats; Support remote interface to send measurement results to cloud server or remote pathology platform in real time.

[0008] A tissue block measurement method based on digital image processing and computer vision is also provided, comprising the following steps: S1, data acquisition, obtaining image data of tissue blocks on the shooting platform; S2, image preprocessing, performing preprocessing operations on the collected image data, including denoising, distortion correction, and white balance; S3, image segmentation and measurement, identifying the outline of the tissue block in the processed image data, and measuring the length, width, area and perimeter of the tissue block; S4, manage the data measured by S3, associate the measurement results with the patient information and store them in the hospital information system; S5. Display the calculation results in real time or generate a report and output it to the designated terminal.

[0009] Preferably, improved calibration and distortion correction are adopted in S1, and the steps are as follows: A1. Place a calibration pattern with multi-scale known sizes on the shooting platform, and use a high-precision camera to calibrate at multiple angles and distances to obtain more comprehensive camera intrinsic and extrinsic parameters. A2. Use multi-scale calibration patterns to achieve higher-precision lens distortion correction, and use a dynamic pixel-to-physical distance conversion algorithm to adjust measurement parameters in real time, thereby significantly improving the accuracy and reliability of tissue block measurements.

[0010] Preferably, calibration and distortion correction use a dual correction model, specifically including: B1. First, the camera's intrinsic parameters (such as focal length, principal point position, radial distortion coefficient) are independently corrected, and the image data of the known size calibration pattern at multiple angles and distances are used to estimate the camera's intrinsic parameters through nonlinear least squares method or other optimization algorithms; B2. Then, the camera's external parameters (such as the position and attitude relationship between the camera and the calibration plate) are independently corrected, and the camera's external parameters are adjusted through an iterative optimization algorithm using the same set of calibration pattern data; B3. After the correction of internal and external parameters is completed, the dual correction model is combined and the estimated values ​​of the camera's internal and external parameters are further accurately adjusted through multiple iterative optimization processes to minimize the measurement error and achieve high-precision pixel-to-physical distance conversion.

[0011] Preferably, the measurement included in S3 comprises the following steps: C1. Identify the contour boundary of the tissue block and extract the contour point set; C2. Calculate the minimum circumscribed rectangle or minimum circumscribed polygon based on the contour point set to obtain the length, width and area of ​​the tissue block; C3. Perform secondary calibration on the final measurement value, the original shooting parameters and the calibration parameters to obtain accurate measurement results.

[0012] Preferably, identifying the contour boundary of the tissue block includes: D1. Edge detection is performed using the Canny operator or the Sobol operator; D2. Use Gabor filtering or Gaussian filtering to further remove noise; D3. Perform morphological processing such as closing and opening operations on the segmentation results to improve the recognition accuracy of the edges of irregular tissue blocks.

[0013] Preferably, S4 includes: E1. Binding the measurement result with the corresponding pathology number or patient ID; E2. Synchronize to the hospital's LIS (laboratory information system) or PACS system to facilitate subsequent pathological diagnosis and scientific research statistics; E3. Generate timestamp and operation records for each measurement to achieve full traceability.

[0014] Therefore, the present invention adopts the above-mentioned tissue block measurement method and system based on digital image processing and computer vision, which has the following beneficial effects: (1) Independently correct internal and external parameters to ensure the accuracy of camera parameters and reduce measurement deviations caused by parameter errors; (2) Angle and multi-distance calibration methods to adapt to different shooting conditions and improve the flexibility of the system; (3) Through multiple iterative optimizations, the parameters are further accurately corrected to achieve high-precision pixel-to-physical distance conversion, significantly reducing measurement errors.

[0015] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a flow chart of the tissue block measurement method based on digital image processing and computer vision of the present invention; Figure 2 This is a schematic diagram of the structure of a tissue block measurement system according to an embodiment of the present invention; Figure 3 Schematic diagram of the flow of the image segmentation and measurement module according to an embodiment of the present invention. DETAILED DESCRIPTION

[0017] The following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] See also Figure 1-3 , a tissue block measurement system based on digital image processing and computer vision, comprising: The data acquisition module includes a shooting platform, which is used to obtain image data of the tissue block on the shooting platform.

[0019] (1) Shooting platform; A dedicated shooting platform is configured in the operating room or pathology department. The platform is equipped with a calibrated scale reference plate or calibration pattern (such as a chessboard) and a bracket for fixing an industrial camera or a high-resolution camera. The platform must have a waterproof and anti-fouling design to adapt to routine cleaning and disinfection operations in a medical environment.

[0020] (2) Image acquisition; Place the removed tissue block flat on the designated area of ​​the shooting platform, and the system can shoot the tissue block by automatic focus or manual focus; The camera is calibrated before shooting or regularly to correct lens distortion and record the conversion relationship between pixels and actual physical distance. Users (such as pathologists) can view the real-time view on the touch screen or software interface and adjust the position of the tissue block so that it enters the lens field of view completely and clearly.

[0021] The image preprocessing module is used to perform denoising, distortion correction, and white balance preprocessing operations on image data.

[0022] Image segmentation and measurement module, used to identify the outline of tissue blocks and calculate the length, width, area and perimeter of tissue blocks.

[0023] The data management module is used to associate the calculation results with the patient information and store them in the hospital information system.

[0024] The result output module is used to display the calculation results in real time or generate reports and output them to the specified terminal. It includes: The visual interface is used to display the measured length, width, area, perimeter and other values; the measurement results (such as length, width, area, perimeter and other values) are displayed instantly on the device screen or the computer of the pathology department; the effect diagram of the segmentation contour superimposed on the original image can be displayed for quick verification by doctors or technicians.

[0025] Automatically generate measurement reports in PDF or other formats; the system can generate PDF reports or other document formats with one click. The report content includes: Tissue block image (including automatically marked length and width); size measurement results and units; corresponding patient information, pathology number, acquisition time, operator, etc.; reports can be printed or archived electronically to facilitate remote consultation or scientific research cooperation within and outside the hospital.

[0026] Support remote interface (or cloud) to send measurement results to cloud server or remote pathology platform in real time. If remote interface or cloud service is configured, measurement data and images can be uploaded to the cloud in real time; subsequent analysis, comparison or data mining can be performed in the cloud, and support can be provided for remote pathology diagnosis.

[0027] A tissue block measurement method based on digital image processing and computer vision, based on the above-mentioned system, comprises the following steps: S1. Data acquisition: obtaining image data of the tissue block on the shooting platform.

[0028] S2: Image preprocessing: performing preprocessing operations on the collected image data, including denoising, distortion correction, and white balance.

[0029] (1) Denoising and enhancement; Because the lighting in the medical environment may be uneven and there may be interference such as blood or mucus on the tissue surface, the original image obtained may contain noise and uneven brightness; Gaussian filtering or median filtering is used to remove random noise while preserving the edge details of the tissue block as much as possible.

[0030] Gaussian filtering: ; Among them, x, y represent the horizontal and vertical distances from the filter center within the filter window (the unit can be pixels).

[0031] Represents the standard deviation of the Gaussian function and controls the degree of smoothing. The larger it is, the more obvious the smoothing effect of the filter will be.

[0032] The kernel function corresponding to the formula performs convolution operation with the original image to obtain the filtered result.

[0033] Median filter: ; Here, Ω represents the neighborhood set centered at (x, y), for example, Ω={−k,…,k}×{−k,…,k}.

[0034] Median It means taking the median value of this group of neighborhood pixel values ​​(in an even-sized window, usually taking the average of the two middle numbers or rounding down).

[0035] (2) Distortion correction; Combined with the calibration process in the data acquisition module, an improved calibration and distortion correction strategy is adopted to improve the measurement accuracy. Specifically, it includes: A. Use of multi-scale calibration patterns: Place calibration patterns with multi-scale known sizes on the imaging platform, such as checkerboards or concentric circular calibration plates of different sizes. These calibration patterns cover multiple positions and size ranges of the tissue block imaging area, ensuring effective calibration at different imaging angles and distances.

[0036] B. Multi-angle and multi-distance calibration process: The multi-scale calibration patterns are captured by a high-precision camera at multiple angles (such as 0°, 45°, 90°, etc.) and multiple distances (such as 10cm, 20cm, 30cm, etc.). Using these diverse calibration images, advanced computer vision algorithms (such as the camera calibration function in OpenCV) are used to accurately extract the camera's intrinsic parameters (such as focal length, principal point position) and extrinsic parameters (such as the position and posture of the camera relative to the calibration plate).

[0037] C. Dynamic pixel-physical distance conversion algorithm: Based on the above calibration results, a dynamic pixel-to-physical distance conversion algorithm was developed. This algorithm can adjust the measurement parameters in real time according to the camera position and angle during actual shooting, and achieve high-precision conversion from pixel coordinates to physical dimensions. Compared with the traditional fixed-ratio conversion, this algorithm can adapt to more shooting changes and significantly improve the accuracy and reliability of measurement.

[0038] D. Optimize the distortion correction model: The distortion correction of this embodiment not only relies on a single distortion model, but also combines multi-scale and multi-angle calibration data and adopts a more complex distortion correction model (such as a mixed model of radial distortion and tangential distortion) to effectively reduce the impact of lens distortion on the measurement results and ensure the size measurement accuracy of the tissue block.

[0039] (3) White balance and brightness adjustment; According to the difference between the shooting light source and the actual color of the tissue block, the white balance is adjusted to make the color restoration of the image more realistic; At the same time, the brightness or contrast can be fine-tuned according to the overall brightness histogram of the image to improve the image visualization effect.

[0040] (4) Dual correction model and multiple iterative optimization strategy; In order to further improve the accuracy of tissue block measurement, this embodiment adopts a dual correction model in the image preprocessing module and combines multiple iterative optimization strategies, which specifically include the following steps: 4.1 Double Correction Model; 4.1.1 Internal reference calibration; First, the camera's internal parameters (intrinsic parameters), such as focal length (f_x, f_y), principal point position (c_x, c_y), and radial distortion coefficient (k_1, k_2, k_3), are calibrated using multi-scale calibration patterns of known size (such as a chessboard of various sizes). The specific process is as follows: 1. Multi-angle and multi-distance acquisition: Place checkerboard calibration patterns of different sizes on the shooting platform, and use a high-precision camera to capture images from multiple angles (such as 0°, 45°, 90°) and distances (such as 10cm, 20cm, 30cm).

[0041] 2. Feature point extraction and matching: Use computer vision algorithms (such as the corner detection function in OpenCV) to extract the corner points in each calibration image and match them with the geometric features of the actual calibration pattern.

[0042] 3. Internal parameter estimation: Using the extracted feature points, the camera's intrinsic parameters are estimated through a nonlinear least squares method (such as the Levenberg-Marquardt algorithm) to minimize the error between the image coordinates and the actual physical coordinates.

[0043] 4.1.2 External reference calibration; After completing the internal parameter calibration, the camera's external parameters (extrinsic parameters), that is, the position and attitude relationship between the camera and the calibration plate, are independently corrected. The specific steps are as follows: 1. Pose Estimation: The PnP (Perspective-n-Point) algorithm is used to estimate the camera's posture (rotation matrix R and translation vector T) relative to the calibration plate using the image data of the known calibration pattern at different positions and angles.

[0044] 2. External parameter optimization: Iterative optimization algorithms (such as the Gauss-Newton method) are used to further accurately estimate extrinsic parameters based on multiple sets of calibration image data to ensure that the geometric relationship between the camera and the calibration plate is accurately described.

[0045] 4.2 Multiple Iteration Optimization: In order to further reduce the measurement error, this embodiment introduces multiple iterative optimization processes, specifically including: 1. Initial Estimates: Based on the above double calibration model, preliminary estimates of internal and external parameters were obtained.

[0046] 2. Error calculation and feedback: An error function is calculated using the error between the actual measurement results (eg, the measurement value of a tissue block of known size) and the corrected pixel-to-physical distance conversion value.

[0047] 3. Optimization and adjustment: Through multiple iterations, the estimated values ​​of the internal and external parameters are adjusted to minimize the error function. After each iteration, the pixel-to-physical distance conversion is re-performed to ensure the gradual refinement of the measurement parameters.

[0048] 4. Convergence judgment: Set an error threshold. When the error drops below the preset threshold or reaches the maximum number of iterations, stop the optimization process and determine the final internal and external parameter estimates.

[0049] S3: Image segmentation and measurement: identifying the outline of the tissue block in the processed image data, and measuring the length, width, area and perimeter of the tissue block.

[0050] (1) Image segmentation method; ①Threshold segmentation and morphological operation (conventional method); Use adaptive threshold or fixed threshold to binarize the image and preliminarily distinguish tissue blocks from background; perform morphological processing such as opening and closing operations on the segmentation results to remove small noise points and fill micro holes; use edge detection (such as Canny operator) to further refine the extraction of tissue block contours. The specific formula is: ; in, is a non-maximum suppression operation, It is a delayed processing operation. and is the low threshold and the high threshold; ② Deep learning refined segmentation (optional method); The model is pre-trained using neural network structures such as U-Net or Mask R-CNN, and can automatically learn the edge features of various tissue blocks. During online reasoning, the pre-processed image is input into the neural network to obtain pixel-level segmentation results, which is suitable for scenes with extremely irregular shapes or complex backgrounds.

[0051] (2) Contour extraction and measurement; Obtain the boundary contour of the tissue block from the segmentation result; calculate the minimum circumscribed rectangle of the contour, whose length and width are regarded as the length and width of the tissue block; further obtain the area (Area) and perimeter (Perimeter) of the tissue block according to the distribution of contour pixels; Combined with the camera calibration information or scale reference, the conversion from pixels to physical distance is completed, and the actual measurement value such as accurate millimeters or centimeters is output.

[0052] (3) Secondary correction; Taking into account factors such as the shooting angle or slight warping of the tissue block, the measurement submodule can perform secondary correction based on the calibration parameters; for extremely irregular shapes, polygon fitting or contour point set interpolation can also be used to further improve the measurement accuracy.

[0053] S4. Manage the data calculated in S3, associate the calculation results with the patient information and store them in the hospital information system.

[0054] (1) Information association and storage; After the measurement is completed, the system binds the dimensional data (length, width, area, circumference, etc.) to the patient ID or pathology number; generates a unique identifier for each tissue block measurement, and records key metadata such as timestamp, operator, and instrument information to achieve traceable management.

[0055] (2) System docking; Upload the measurement results to the hospital's LIS (laboratory information system) or PACS (picture archiving and communication system) through a network interface or API; if the hospital already has an electronic medical record system (EMR) or other scientific research database, standardized data formats (such as HL7, DICOM, etc.) can also be used for docking and integration.

[0056] S5. Display the calculation results in real time or generate a report and output it to the designated terminal.

[0057] Therefore, the present invention adopts the above-mentioned tissue block measurement method and system based on digital image processing and computer vision, which can improve the correction accuracy, enhance the system adaptability and reduce the measurement error.

[0058] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. A tissue block measurement system based on digital image processing and computer vision, characterized in that: include: A data acquisition module, including a shooting platform, is used to obtain image data of the tissue block on the shooting platform; Image preprocessing module, used for performing denoising, distortion correction, and white balance preprocessing operations on image data; Image segmentation and measurement module, used to identify the outline of tissue blocks and calculate the length, width, area and perimeter of tissue blocks; A data management module is used to associate the calculation results with the patient information and store them in the hospital information system; The result output module is used to display the calculation results in real time or generate reports and output them to the terminal.

2. The tissue block measurement system based on digital image processing and computer vision according to claim 1, characterized in that: The result output modules include: Visual interface for displaying measured length, width, area and perimeter; Automatically generate measurement reports; Support remote interface to send measurement results to cloud server or remote pathology platform in real time.

3. A tissue block measurement method based on digital image processing and computer vision, using the tissue block measurement system based on digital image processing and computer vision as described in any one of claims 1-2, characterized in that: The following steps are involved: S1, data acquisition, obtaining image data of tissue blocks on the shooting platform; S2, image preprocessing, performing preprocessing operations on the collected image data, including denoising, distortion correction, and white balance; S3, image segmentation and measurement, identifying the outline of the tissue block in the processed image data, and measuring the length, width, area and perimeter of the tissue block; S4. Manage the measured data, associate the measurement results with the patient information and store them in the hospital information system; S5. Display the calculation results in real time or generate a report and output it to the terminal.

4. The tissue block measurement method based on digital image processing and computer vision according to claim 3, characterized in that: S1 uses improved calibration and distortion correction, the steps are as follows: A1. Place a calibration pattern with multi-scale known sizes on the shooting platform, perform multi-angle and multi-distance calibration through the camera, and obtain the camera's intrinsic and extrinsic parameters; A2. Use multi-scale calibration patterns to achieve lens distortion correction, and adjust measurement parameters in real time through a dynamic pixel-to-physical distance conversion algorithm.

5. The tissue block measurement method based on digital image processing and computer vision according to claim 4, characterized in that: Calibration and distortion correction use a dual correction model, including: B1. Independently correct the camera's intrinsic parameters, use the image data of the known size calibration pattern at multiple angles and distances, and estimate the camera's intrinsic parameters through an optimization algorithm; B2. Independently correct the camera's extrinsic parameters, and use the same set of calibration pattern data to adjust the camera's extrinsic parameters through an iterative optimization algorithm; B3. After the correction of internal and external parameters is completed, the dual correction model is combined to adjust the estimated values ​​of the camera's internal and external parameters through multiple iterative optimization processes to achieve pixel-to-physical distance conversion.

6. The tissue block measurement method based on digital image processing and computer vision according to claim 5, characterized in that: The measurements included in S3 include the following steps: C1. Identify the contour boundary of the tissue block and extract the contour point set; C2. Calculate the minimum circumscribed rectangle or minimum circumscribed polygon based on the contour point set to obtain the length, width and area of ​​the tissue block; C3. Perform secondary calibration on the final measured value, the original shooting parameters and the calibration parameters to obtain the measurement result.

7. The tissue block measurement method based on digital image processing and computer vision according to claim 6, characterized in that: Identifying the outline boundary of the tissue block includes: D1. Edge detection is performed using the Canny operator or the Sobol operator; D2. Use Gabor filtering or Gaussian filtering to remove noise; D3. Perform closing and opening morphological processing on the segmentation results to improve the recognition accuracy of the edges of irregular tissue blocks.

8. The tissue block measurement method based on digital image processing and computer vision according to claim 7, characterized in that S4 include: E1. Bind the measurement results to the corresponding pathology number or patient ID; E2. Synchronize to the hospital laboratory information system LIS or PACS system; E3. Generate timestamp and operation records for each measurement.

Images