Dynamic threshold skin surface point cloud wound positioning segmentation method based on region growth
By adopting a dynamic threshold method based on regional growth in the skin surface point cloud data processing, the problem of inaccurate wound segmentation in the prior art is solved, and the precise positioning and segmentation of irregular and uneven wounds is achieved, which significantly improves the scientificity of the evaluation and the widespread clinical application.
Patent Information
- Application Number
- CN202411885710.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-20
- Publication Date
- 2025-04-22
AI Technical Summary
When processing point cloud data on the skin surface, it is difficult to accurately segment protrusions or depressions, and problems of undersegment and oversegment often occur, affecting the scientific nature of the evaluation and the widespread clinical application.
The dynamic threshold method based on regional growth is adopted, and by flexibly adjusting the threshold value and combining the changes in the angle cosine value of the normal vector, it dynamically adapts to the geometric characteristics of the point cloud, and realizes the precise positioning and segmentation of irregular and uneven wounds.
It significantly improves the accuracy and robustness of wound segmentation, enables accurate identification of wound boundaries and shapes, provides reliable clinical support, and promotes the accuracy of wound assessment.
Smart Images
Figure CN119863484A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of machine vision technology, and more specifically, to a dynamic threshold skin surface point cloud wound positioning and segmentation method based on region growing. Background Art
[0002] Currently, the assessment of skin lesions mainly relies on subjective judgment, lacking standardized qualitative methods and clear grading indicators. This assessment method has significant subjectivity and inconsistency, which may lead to deviations and unpredictable effects. Therefore, an objective and accurate assessment method is urgently needed to solve this problem.
[0003] However, the existing technology has not yet provided a scientific and systematic skin injury grading standard, which affects the scientific nature of injury assessment and the wide range of clinical applications. In particular, when performing wound segmentation and extraction on the skin surface point cloud file, the wound appears as a protrusion or depression in the point cloud. The traditional region growing algorithm often has problems of under-segmentation and over-segmentation when processing such point cloud data, which further highlights the need to develop more advanced assessment and segmentation methods. Summary of the invention
[0004] In view of the shortcomings of the prior art, the present invention aims to solve the defects of the wound part segmentation method in the prior art and proposes a curvature threshold dynamic adjustment method based on region growing and the cosine value of the normal vector angle.
[0005] Through flexible threshold adjustment, this method can solve the over-segmentation problems caused by under-segmentation of protrusions, depressions, different color parts of the wound, and congestion, so that the segmentation results are highly consistent with the actual wound, thereby significantly improving the accuracy and robustness of the segmentation. This point cloud skin wound segmentation method can accurately identify the boundaries and shapes of wounds, providing reliable support for clinical applications and effectively promoting the accuracy of wound assessment.
[0006] The first aspect of the present invention provides a dynamic threshold skin surface point cloud wound location and segmentation method based on region growing, comprising:
[0007] Step 1: Obtain skin surface point cloud data;
[0008] Step 2: Filter the acquired point cloud data;
[0009] Step 3: Randomly place points on the non-wound part, and determine the range and edge of the wound part point cloud after region growing;
[0010] Step 4: Establish a dynamic threshold based on the characteristics of the point cloud in the target area;
[0011] Step 5: Perform region growing processing on the surface point cloud data of the wound area to obtain the segmentation result.
[0012] The second aspect of the present invention provides a dynamic threshold skin surface point cloud wound positioning and segmentation system based on region growing, comprising: an acquisition unit for acquiring skin surface point cloud data; a filtering unit for filtering the acquired point cloud data; a growth unit for randomly landing points on the non-wound part, and determining the range and edge of the wound part point cloud after performing region growing; a setting unit for establishing a dynamic threshold based on the characteristics of the target area point cloud; and a segmentation unit for performing region growing processing on the surface point cloud data of the wound range to obtain a segmentation result.
[0013] The third aspect of the present invention provides an electronic device for dynamic threshold skin surface point cloud wound localization and segmentation based on region growing, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the above-mentioned dynamic threshold skin surface point cloud wound localization and segmentation method based on region growing is implemented.
[0014] A fourth aspect of the present invention provides a computer-readable storage medium storing a computer program for executing the above-mentioned dynamic threshold skin surface point cloud wound localization and segmentation method based on region growing.
[0015] The present invention produces the following technical effects:
[0016] The dynamic threshold segmentation method based on region growing of the present invention shows excellent results in processing the point cloud segmentation of wounds with uneven surfaces and different colors. The method dynamically adapts to the geometric characteristics of the point cloud by adjusting the threshold in real time according to the change of the cosine value between the normal vectors, thereby effectively processing the point cloud data with protrusions or depressions and color differences. This flexible threshold adjustment mechanism significantly reduces the risk of over-segmentation and under-segmentation. Therefore, the segmentation algorithm is particularly suitable for the precise positioning and segmentation of wounds on irregular and uneven skin surfaces. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] Figure 1 This is a flowchart of skin surface point cloud wound positioning and segmentation in an embodiment of the present application;
[0018] Figure 2 is a flow chart of step 2 of the embodiment of the present application;
[0019] Figure 3 is a flow chart of step 3 of the embodiment of the present application;
[0020] Figure 4 is a flow chart of step 4 of the embodiment of the present application;
[0021] Figure 5 is a flow chart of step 5 of the embodiment of the present application;
[0022] Figure 6 This is the flow chart of step 6 of the embodiment of the present application. DETAILED DESCRIPTION
[0023] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0024] Figure 1 The process of skin surface point cloud wound location segmentation based on dynamic threshold region growing proposed in the embodiment of the present application is shown, which includes the following steps in sequence:
[0025] S1: The wound is scanned using laser scanning to obtain point cloud data of the skin surface.
[0026] S2: Filter the acquired point cloud data to remove outlier point clouds and reduce point cloud density, reduce noise impact and reduce computational time complexity.
[0027] S3: Randomly place points on the non-wound part, and perform regional growth with the cosine value of the normal vector angle as the growth condition to determine the range and edge of the point cloud of the wound part.
[0028] First, the point cloud is divided into regions through the octree to facilitate point cloud search and reduce the complexity of calculation time. In the initial segmentation area, seed points are randomly placed and the straight-line distance between the seed points and the neighboring points is calculated. The cosine value and curvature value of the angle between the normal vector and the points within the distance range are calculated.
[0029] By calculating the cosine value of the normal vector angle and using the cosine value as the region growth judgment condition, the wound area range and boundary are initially segmented to eliminate the influence of factors such as congestion on the subsequent wound segmentation accuracy.
[0030] S4: Establish a dynamic threshold based on the characteristics of the point cloud in the target area as a constraint.
[0031] And calculate the color threshold of points within the distance range.
[0032] S5: Performing region growing processing on the surface point cloud data of the wound range to obtain a segmentation result.
[0033] S6: Use the segmentation result to perform three-dimensional reconstruction of the wound point cloud.
[0034] S7: Evaluate the segmentation results to ensure the accuracy and effectiveness of the segmentation.
[0035] Figure 2The process of preprocessing the acquired point cloud is shown in the following steps:
[0036] A1: Obtain original skin surface point cloud data.
[0037] A2: Use the Statistical Outlier Removal algorithm to filter the outlier point cloud.
[0038] A3: Use voxel filtering to reduce the point cloud density and the total number of point clouds, and reduce the computational time complexity of the final traversal.
[0039] A4: Extract pre-processed point cloud data to prepare for subsequent analysis.
[0040] Figure 3 The process of determining the range and boundary of the wound area point cloud is shown. The steps include:
[0041] B1: Extract preprocessed point cloud data.
[0042] B2: Seed placement point in non-wound area.
[0043] B3: Set the cosine value of the angle between the point cloud normal vector and the point cloud as the threshold of the region growth constraint.
[0044] B4: Perform single-seed region growing iterations to determine the scope and edge of the wound point cloud.
[0045] B5: Preliminary segmentation determines the boundary and range of the wound point cloud.
[0046] Figure 4 The process of dynamic threshold construction is described. The process includes:
[0047] C1: Traverse each point and construct a local plane p based on each point p and the neighboring point k.
[0048] C2: Use PCL::Normal to estimate the normal vector for each point.
[0049] C3: Use the covariance matrix to calculate the point cloud curvature based on the estimated normal vector.
[0050] C4: Select point cloud p and calculate the cosine value of the angle α of its neighborhood normal vector.
[0051] C5: Repeat step 4 for each point and calculate the cosine value cosα of the normal vector angle of all point clouds.
[0052] C6: Set dynamic thresholds according to the cosα value to ensure segmentation flexibility and adaptability. By analyzing the changes in cosα, the curvature threshold is dynamically adjusted to improve the positioning accuracy and segmentation effect of the wound site.
[0053] Figure 5 The process of dynamic threshold region growing segmentation is presented, including:
[0054] D1: Select multiple points as seed points in the initial segmentation area.
[0055] D2: For each adjacent point cloud, determine whether the point cloud belongs to the target area where the seed is located according to the predefined threshold standard. Specifically: determine whether the cosine value of the angle between the seed point and the normal vector of the point within the range is greater than the set threshold, so as to choose to use color similarity or curvature similarity to divide the points into regions. This method can determine different color areas on the same wound as the same area during segmentation.
[0056] D3: Add the adjacent point clouds belonging to the target area to the seed point set to form a new seed point set.
[0057] D4: Repeat the above steps for the new seed set points and perform region growing iteratively.
[0058] D5: After the iteration, the final segmentation result is output to achieve accurate wound segmentation. Figure 6 The reconstruction process of the wound point cloud model after segmentation is shown. The reconstruction process includes:
[0059] F1: Point cloud data extracted from region segmentation.
[0060] F2: Initialize the point cloud, taking a point as the initial starting point.
[0061] F3: Perform greedy triangle calculation on the extracted wound point cloud to form a triangular patch.
[0062] F4: Generate a 3D model of the wound surface using continuous triangular patches obtained through greedy triangulation to achieve wound visualization and analysis.
[0063] This embodiment provides a dynamic threshold skin surface point cloud wound positioning and segmentation system based on region growing, and applies the dynamic threshold skin surface point cloud wound positioning and segmentation method based on region growing described in the above embodiment, including:
[0064] The acquisition unit is used to acquire skin surface point cloud data.
[0065] The filtering unit is used to perform filtering processing on the acquired point cloud data.
[0066] The growth unit is used to randomly place points on the non-wound part and determine the range and edge of the point cloud of the wound part after regional growth.
[0067] The value setting unit is used to establish a dynamic threshold according to the characteristics of the point cloud in the target area.
[0068] The segmentation unit is used to perform region growing processing on the surface point cloud data of the wound range to obtain a segmentation result.
[0069] The embodiment of the present application also discloses an electronic device for positioning and segmenting wounds based on dynamic threshold skin surface point cloud based on region growing, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method for positioning and segmenting wounds based on dynamic threshold skin surface point cloud based on region growing when executing the computer program, wherein the memory may include a memory, such as a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk memory, etc.; the processor, the network interface, and the memory are interconnected through an internal bus, and the internal bus may be an industrial standard architecture bus, a peripheral component interconnection standard bus, an extended industrial standard architecture bus, etc., and the bus may be divided into an address bus, a data bus, a control bus, etc. The memory is used to store programs, and specifically, the program may include a program code, and the program code includes computer operation instructions. The memory may include a memory and a non-volatile memory, and provide instructions and data to the processor.
[0070] The embodiment of the present application also discloses a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method for locating and segmenting a skin surface point cloud wound based on dynamic threshold region growth are implemented. Specifically, the computer-readable storage medium includes, but is not limited to, for example, volatile memory and / or non-volatile memory. The volatile memory may include a random access memory (RAM) and / or a cache memory, etc. The non-volatile memory may include a read-only memory (ROM), a hard disk, a flash memory, an optical disk, a magnetic disk, etc.
[0071] 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 the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.
Claims
1. A dynamic threshold skin surface point cloud wound location and segmentation method based on region growing, characterized in that: The specific steps are: Step 1: Obtain skin surface point cloud data; Step 2: Filter the acquired point cloud data; Step 3: Randomly place points on the non-wound part, and determine the range and edge of the wound part point cloud after region growing; Step 4: Establish a dynamic threshold based on the characteristics of the point cloud in the target area; Step 5: Perform region growing processing on the surface point cloud data of the wound area to obtain the segmentation result.
2. The method for skin surface point cloud wound location and segmentation based on dynamic threshold and region growing according to claim 1, characterized in that: The step 2 specifically includes: Step 2.1: Obtain original skin surface point cloud data; Step 2.2: Use the statistical outlier removal algorithm to filter the outlier point cloud; Step 2.3: Use voxel filtering to reduce the point cloud density, the total number of points, and the computational time complexity of the final traversal.
3. The method for skin surface point cloud wound location and segmentation based on dynamic threshold and region growing according to claim 1, characterized in that: The step 3 specifically includes: Step 3.1: Extract preprocessed point cloud data; Step 3.2: seeding points in non-wound areas; Step 3.3: Set the cosine value of the angle between the point cloud normal vector and the point cloud as the threshold of the region growth constraint condition; Step 3.4: Perform single seed region growing iteration to determine the scope and edge of the wound point cloud; Step 3.5: Preliminary segmentation determines the boundary and range of the wound point cloud.
4. The method for skin surface point cloud wound location and segmentation based on dynamic threshold and region growing according to claim 1, characterized in that: The step 4 specifically includes: Step 4.1: Traverse each point and construct a local plane based on each point and its neighboring points; Step 4.2: Use PCL::Normal to estimate the normal vector of each point; Step 4.3: Use the covariance matrix to calculate the point cloud curvature based on the estimated normal vector; Step 4.4: Select the cosine value of the angle between the normal vectors of each point in the point cloud computing neighborhood; Step 4.5: Repeat step 4 for all points until the cosine values of the normal vector angles of all point clouds are calculated; Step 4.6: Set a dynamic threshold according to the cosine value.
5. The method for skin surface point cloud wound location and segmentation based on dynamic threshold and region growing according to claim 1, characterized in that: The step 5 specifically includes: Step 5.1: Select multiple points as seed points in the preliminary segmentation area; Step 5.2: For each adjacent point cloud, determine whether the point cloud belongs to the target area where the seed is located according to the predefined threshold standard; Step 5.3: Add the adjacent point clouds belonging to the target area to the seed point set to form a new seed point set; Step 5.4: Repeat the above steps for the new seed set points and iterate the region growing; Step 5.5: After the iteration, output the final segmentation result.
6. The method for skin surface point cloud wound location and segmentation based on dynamic threshold and region growing according to claim 1, characterized in that: Also includes: Step 6: Use the segmentation results to perform three-dimensional reconstruction of the wound point cloud; Step 7: Evaluate the segmentation results to ensure the accuracy and effectiveness of the segmentation.
7. The method for skin surface point cloud wound location and segmentation based on dynamic threshold and region growing according to claim 6, characterized in that: The step 6 specifically includes: Step 6.1: Point cloud data extracted from region segmentation; Step 6.2: Initialize the point cloud, taking a point as the initial starting point; Step 6.3: Perform greedy triangle calculation on the extracted wound point cloud to form a triangular patch; Step 6.4: Generate a three-dimensional model of the wound surface using the continuous triangular patches obtained by greedy triangulation to achieve wound visualization and analysis.
8. A dynamic threshold skin surface point cloud wound location and segmentation system based on region growing, characterized in that: include: An acquisition unit, used for acquiring skin surface point cloud data; A filtering unit, used for filtering the acquired point cloud data; The growth unit is used to randomly place points on the non-wound part and determine the range and edge of the point cloud of the wound part after regional growth; A value setting unit is used to establish a dynamic threshold value according to the characteristics of the point cloud of the target area; The segmentation unit is used to perform region growing processing on the surface point cloud data of the wound range to obtain a segmentation result.
9. An electronic device for skin surface point cloud wound location and segmentation based on dynamic threshold region growing, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the dynamic threshold skin surface point cloud wound localization and segmentation method based on region growing as described in any one of claims 1 to 8 above is implemented.
10. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and the computer program is used to execute the dynamic threshold skin surface point cloud wound positioning and segmentation method based on region growing as described in any one of claims 1 to 8.