Method, apparatus, and storage medium for automatically identifying and locating perforating vessels.

Abstract

The present invention relates to a method, device, and storage medium for automatically identifying and locating perforator vessels, which includes the following steps: determining a scanning range in the scanning direction from the target perforator, and extracting 3D image data within the scanning range; performing dimensionality reduction processing on the 3D image data to obtain slice-segmented image data; using the slice-segmented image data as the smallest analysis unit, screening the signal points of perforator vessels using a signal-to-noise ratio dynamic threshold method; 3D reconstructing the signal points of the connected perforator vessels in the 3D image data, and screening candidate perforator vessels based on the total score of the signal-to-noise ratio after reconstruction; and identifying the candidate perforator vessels based on their relative position, angle, and depth to the superficial layer and deep muscle layer. The present invention can realize automatic and fast location identification of perforator vessels.

Description

Technical Field

[0001] The present invention relates to the technical field of medical image processing, and particularly to a method for automatically identifying and locating perforator blood vessels.

Background Art

[0002] The perforator flap transplantation surgery is at the forefront of the microsurgery field and is an indispensable technique for achieving the best repair effect in the reconstruction of body surface organs (such as breast reconstruction), trauma or tumor defect repair, and minimizing the damage to the donor area.

[0003] A flap is a composite tissue piece including skin and subcutaneous tissue, and the tissue piece contains its own nutrient blood vessels. Usually, the flap moves from one place to another in the body to repair the defect caused by the above reasons, and such a surgical process is called a flap transplantation surgery. A perforator blood vessel is a blood vessel that penetrates the skin and its corresponding subcutaneous tissue, and its function is to supply blood to the penetrated skin. The characteristic of a perforator flap is that the perforator blood vessels supplying nutrition to the flap are dissected completely to the skin penetration point, and the original main trunk blood vessels are retained, and other tissues such as muscles supplied by the main trunk blood vessels are not affected. The conventional flap resection technique does not dissect to the skin penetration point and directly disconnects the blood vessels at the main trunk, so the damage is large.

[0004] The main technical difficulty of the perforator flap is that the variation of the perforator blood vessels is large, and there are significant differences not only between different human bodies but also between different positions on both sides of the same body. And the caliber of the skin penetration point of the perforator blood vessels is very thin (0.5 - 1 mm). When its specific position is unknown, it is easy to be damaged, and the damage means surgical failure, so the surgical risk is high and the stress of the surgeon during the operation is great.

[0005] By visualizing the location of perforating vessels' skin penetration points preoperatively, surgeons can know the location of these points before surgery, significantly reducing the risk and time required for intraoperative dissection, thereby improving surgical success rates. Tomographic medical imaging (e.g., enhanced CT angiography, CTA, enhanced MRI angiography, MRA) is a common method. Tomographic techniques for locating perforating vessels are relatively mature, offering high accuracy and allowing for confirmation of vessel location information from different angles. Subsequently, volume rendering (VR) technology is used to mark the skin penetration points of the perforating vessels on the patient's body surface.

[0006] However, a major limitation of the above method is that the image analysis and localization work is completed by the surgeons themselves on the surgical team, and domestic radiologists cannot perform this task due to limitations in their understanding of anatomy or other reasons, and few radiologists are proficient in this technique. The time required for tomographic image analysis and marking of perforating vessels is usually around 45 minutes, which is very draining for the surgeon and involves repetitive, mechanical work. [Overview of the Initiative]

[0007] The technical problem that the present invention aims to solve is to provide a method for automatically identifying and locating perforating vessels, which enables the automatic and rapid locating of perforating vessels.

[0008] The present invention provides a method for automatically identifying and locating perforating vessels, comprising the following steps: From the target perforating branch, the scanning range in the scanning direction is determined, and 3D image data within the scanning range is extracted; The aforementioned three-dimensional image data is subjected to dimensionality reduction processing to obtain layered image data; The aforementioned stratified image data is used as the minimum analysis unit, and signal points of perforating vessels are screened using the signal-to-noise ratio dynamic threshold method; The signal points of the connected perforating vessels in the 3D image data are reconstructed in 3D, and candidate perforating vessels are screened using the total signal-to-noise ratio score after reconstruction; The perforating vessels are identified based on their relative position, inclination, and depth relative to the epidermal and deep muscular layers.

[0009] The process of obtaining layered image data by performing dimensionality reduction on the aforementioned three-dimensional image data specifically includes the following: By merging n consecutive images in 3D image data, a signal-enhanced 2D image data is obtained; In 2D image data, the boundaries of the skin surface and the boundaries between subcutaneous fat and muscle are marked, and the subcutaneous fat region enclosed by these two boundaries is defined as the target analysis area; The analysis region of the target is stratified to obtain multiple consecutive stratified image data of subcutaneous depth. When merging n consecutive images in the aforementioned 3D image data, the pixel value of the point with the smallest pixel value at the same position is taken as the pixel value of the merged image.

[0010] The coordinates of the boundary between the air and the skin surface are obtained by a machine learning algorithm, and the coordinates of the boundary between the subcutaneous fat and muscle are obtained from the rate of change of the average pixel value.

[0011] Screening signal points in perforating vessels using the aforementioned stratified image data as the minimum analysis unit and the signal-to-noise ratio dynamic threshold method specifically includes the following: The same subcutaneous depth stratified image data is converted into a two-dimensional array, the minimum or maximum pixel value at several pixel points at different depths in the two-dimensional array at similar skin locations is used as a candidate vascular signal point, and the minimum or maximum pixel value at a nearby location is used as a background reference value, and the relative signal-to-noise ratio of the candidate vascular signal points is calculated; A set of points is selected where the pixel value is smaller than the pixel threshold. From this set, points where the relative signal-to-noise ratio exceeds the signal-to-noise ratio threshold are selected as signal points for perforating vessels, and the corresponding horizontal coordinates are marked.

[0012] When calculating the relative signal-to-noise ratio of the candidate vascular signal points mentioned above, it is calculated using ((X4+X5 / 2)*(X2+X3+X4+X5)) / (X0+X1), where X n X0 is the average value of pixels located n pixels away from the candidate blood vessel signal point, and X0 is the pixel value of the candidate blood vessel signal point.

[0013] Specifically, the 3D reconstruction of the signal points of the connected perforating vessels in the aforementioned 3D image data is performed as follows: The transverse coordinates corresponding to the signal points of the perforating vessels are projected onto the two-dimensional image data, points adjacent to the signal points of the perforating vessels in the two-dimensional image data are merged, and the sum is calculated to determine the continuity of the signal points of the merged perforating vessels, the spatial propagation method, and the signal-to-noise ratio; Two-dimensional image data is merged to form a three-dimensional space, the merged points in the two-dimensional image data are placed into the three-dimensional space, the merged points in the three-dimensional space are merged with points adjacent to the merged points in the two-dimensional image data, and the sum is calculated to determine the signal-to-noise ratio after merging.

[0014] After step (5) above, the total score of each perforating vessel is calculated based on the signal-to-noise ratio and vascular mass, and the perforating vessel with the highest total score is designated as the final perforating vessel.

[0015] The technical solution used to solve the technical problem of the present invention is a computer system comprising a storage device and a processor, wherein a computer program is stored in the storage device, and when the computer program is executed by the processor, the processor performs the steps of a method for automatically identifying and locating the perforating vessels.

[0016] The technical solution used to solve the technical problem of the present invention is a readable storage medium, in which an executable computer program is stored. When the executable computer program is executed by a processor, the processor executes the steps of the method for automatically identifying and positioning the above-mentioned perforating branch blood vessels.

Advantages of the Invention

[0017] By using the above technical solution, the present invention has the following advantages and positive effects compared with the prior art. The present invention reduces the dimension of 3D image data to 2D and searches for signal points of perforating branches in 1D images, so the processing speed is improved, and the operation of positioning perforating branch blood vessels can be completed within 5 minutes. The method of the present invention takes the artificial judgment result of a doctor familiar with the technology of positioning perforating branch blood vessels in tomographic medical images as the gold standard, and the result of automatically positioning perforating branch blood vessels basically matches the gold standard, and the identification rate of dominant perforating branch blood vessels exceeds 99%.

[0018] Hereinafter, in order to more clearly explain the embodiments of the present invention or the technical solutions in the prior art, the drawings necessary for the description of the embodiments or the prior art will be briefly introduced. Of course, the drawings in the following description are only some embodiments of the present invention, and those skilled in the art can also obtain other drawings from these drawings without creative labor.

Brief Description of the Drawings

[0019] [Figure 1] FIG. 1 is a flowchart of the first embodiment of the present invention. [Figure 2] FIG. 2 is a schematic diagram of the analysis range of 2D image data in the first embodiment of the present invention. [Figure 3] FIG. 3 is a schematic diagram of converting 2D image data into strip shape in the first embodiment of the present invention. [Figure 4] FIG. 4 is a schematic diagram of converting 2D image data into 1D image data in the first embodiment of the present invention. [Figure 5]Figure 5 is a schematic diagram of the output result of the single-layer two-dimensional image data in the first embodiment of the present invention. [Figure 6] Figure 6 is a schematic diagram of the result after merging the two-dimensional image data of different layers at continuous points in the three-dimensional space in the first embodiment of the present invention. [Figure 7] Figure 7 is a schematic diagram of the distinction between subcutaneous veins and perforating branch vessels in the first embodiment of the present invention. [Figure 8] Figure 8 is a schematic diagram of the preliminary comparison result between the calculation result of the first embodiment of the present invention and the gold standard. [Figure 9] Figure 9 is a schematic diagram of the human-computer interaction interface of the second embodiment of the present invention.

Embodiments for Carrying out the Invention

[0020] Hereinafter, the present invention will be further described with specific examples. It should be understood that these examples are only used to explain the present invention and do not limit the scope of the present invention. Also, after reading the content of the present invention, those skilled in the art can make various changes and modifications to the present invention, and it should be understood that those equivalent modes are included in the scope of the claims of the present invention.

[0021] Embodiments of the present invention relate to a method for automatically identifying and locating perforating vessels, which can automatically analyze medical imaging data, automatically calculate and identify target perforating vessels, sort them by degree of dominance, and return their location information, and are useful for clinical and scientific research applications requiring preoperative locating of perforating vessels or identification and locating of other perforating vessels. As shown in Figure 1, the method includes the following steps: determining the scanning range in the scanning direction from the target perforating vessel and extracting 3D image data within the scanning range; performing dimensionality reduction processing on the 3D image data to obtain stratified image data; using the stratified image data as the minimum analysis unit, screening signal points of perforating vessels by a signal-to-noise ratio dynamic thresholding method; 3D reconstructing the signal points of the connected perforating vessels in the 3D image data and screening candidate perforating vessels by the sum score of the reconstructed signal-to-noise ratio; determining the perforating vessels from their relative position, inclination, and depth relative to the epidermal and deep muscular layers, and returning the coordinates of the perforating vessels in a 3D coordinate system.

[0022] This method automatically reads and analyzes data from a normalized tomographic image file (e.g., DICOM), automatically identifies the muscle penetration points and the progression paths of perforating vessels within the subcutaneous soft tissue, scores and sorts the perforating vessels based on information such as signal strength and direction of progression, and outputs information on the location of the perforating vessels, their scores, and sorting status in the calculation results. Specifically, it works as follows:

[0023] 1) Read and analyze the data from the normalized tomography file (read a multidimensional array from the original image information that includes the 3D coordinates of the z, y, and x axes, plus the pixel values ​​of those coordinates).

[0024] 2) Limit the analysis range in the z-axis. Limit the scanning range in the z-axis direction from the target perforating branch and extract the corresponding data. Taking the DIEA (deep inferior epigastric artery) as an example, the original image includes the body from the chest to the knee, and the analysis range required for the DIEA vessel is from the navel to the perineum. Therefore, based on the existing features of the image (the navel appears as a depression in the cross-sectional image), first, identify the boundary of the body surface, then identify the location where the depression appears at the boundary of the body surface as the navel, and take the DICOM image from the navel to the pubic bone as the target of analysis.

[0025] 3) After appropriately merging cross-sectional images of adjacent horizontal planes along the z-axis, 2D image data is obtained, and then the 2D image data is analyzed. To enhance the signal of the penetration point, since subcutaneous perforating vessels progress continuously, the penetration signals in adjacent cross-sectional images are generally adjacent in position and, after merging, are basically continuous or fused. The number of horizontal cross-sectional images in the z-axis direction after merging = number of horizontal cross-sectional images within the range taken in the previous step / n (n = number of adjacent merged images). The logic of merging in this step is to use the point with the smallest image value at the same position in the n adjacent images as the pixel value of the merged image. The analysis is carried out over the merged image, and subsequent analysis is performed using one horizontal cross-sectional image (2D) after merging as the unit, thereby converting the 3D image to a 2D image for analysis.

[0026] 4) Limit the analysis range in the 2D image; that is, mark the boundary between air and the skin surface and the boundary between subcutaneous fat and muscle in the 2D image data, and define the region enclosed by these two boundaries as the target analysis region.

[0027] 4.1 Upper boundary of the analysis range: In this process, this is defined as skin, and the boundary between air and the skin surface is marked using the watershed algorithm.

[0028] 4.2 Lower boundary of analysis range: a) The 2D image is cut into strips, the body surface shape is used as the edges, and the image values ​​of 10 pixels in height are read. These values ​​are placed into a sub-image with a height of 10 and a width equal to the total length of the original image's x-axis. The sub-image is reconstructed to make the irregular shape into a rectangle. b) Determination of the lower boundary of the analysis range: Except for the abdominal cavity and the body surface, the areas are similar, mainly dark in color, and the pixel values ​​of the subcutaneous tissue in the target analysis range are relatively high. A threshold is set for the cut-out rectangular sub-image to determine the lower boundary of the analysis range. The average value of the grayscale values ​​is calculated for each of the cut-out rectangular images with a height of 10 pixels. When the proportion of points with high pixel values ​​that are closer to white reaches the threshold, it is determined that the image has already entered the muscle area and reached the lower boundary of the analysis range. Figure 2 shows the analysis range of the 2D image data in this embodiment. In Figure 2, A is the upper boundary of the analysis range, B is the lower boundary of the analysis range, and the area between them is the target analysis region.

[0029] 5) The 2D image within the analysis range is cut into strips, and each strip is analyzed as a 1D array. Using the method in 4.2 a), the 2D image is cut into strips with a height of 10 pixels, as shown in Figure 3. At this time, the 2D image is made into a 2D array, and then converted into a single 1D array, as shown in Figure 4, the data structure of the 1D array is [corresponding value of x-coordinate 1, corresponding value of x-coordinate 2, ...], where the element number in the array corresponds to the x-coordinate, and the x-coordinate corresponding value is the minimum pixel value among the 10 pixel points corresponding to the same x-coordinate (i.e., the point with the strongest signal and closest to black). The signal-to-noise ratio of the "strong signal" point is calculated, and in this embodiment, the signal-to-noise ratio is the ratio of the pixel value of the point itself to the pixel values ​​of the surrounding points. Since the pixel value of blood vessels is low and close to 0 (dark in color in the image), and the pixel value of the surrounding points is high (closer to white in the image), the larger the difference between the signal point and the surrounding points, the better it matches the characteristics of blood vessels, and the higher the score. The logic for calculating the signal-to-noise ratio in this process is ((X4+X5 / 2)*(X2+X3+X4+X5)) / (X0+X1), where X n X0 is the average value of pixels located n pixels away from the signal point of the perforating branch, and X0 is the pixel value of the signal point of the perforating branch.

[0030] 6) From the one-dimensional image data, points with a "strong signal" (points that are closest to black in the figure and have the smallest pixel value) are found. In this embodiment, points with a "strong signal" become the signal points of perforating vessels, and can be found by the following means: Select points with a pixel value of less than 60 to form a set, and from the set, select points whose relative signal-to-noise ratio is greater than the median cutoff of the set as the perforating vessel signal points, and mark the horizontal coordinates corresponding to the perforating vessel signal points. If the distance between the horizontal coordinates corresponding to multiple perforating vessel signal points is too close, they may be appropriately merged.

[0031] Steps 5)-6) above are repeated until each 2D image is divided and layered, and the calculation is completed.

[0032] 7) At locations where there are "strong signal" points in the 2D image, adjacent points are merged, and the signal-to-noise ratios of these points are added together. That is, the transverse coordinates corresponding to the signal points of the perforating vessels are projected onto the 2D image data, points adjacent to the signal points of the perforating vessels in the 2D image data are merged, and the sum of the signal-to-noise ratios of these points is calculated to obtain the continuity, spatial propagation method, and signal-to-noise ratio of the merged signal points of the perforating vessels. Figure 5 is a schematic diagram of the output result of a single-layer 2D image.

[0033] 8) Points in the 2D image from the previous step are placed in 3D, adjacent points are merged, and their signal strengths are summed. Although signals from the same perforating branch appear in different scan cross-sections, their relative positions are adjacent. Therefore, in this step, the 2D images are merged to form a 3D space, the previously identified points are placed in the 3D space to determine their distance, adjacent points are merged, and then their signal strengths are summed (see Figure 6).

[0034] Furthermore, each threshold value in the above process can be calibrated using actual patient imaging data, and after adjustment, the optimal calculation result can be obtained.

[0035] 9) Distinguishing between subcutaneous veins and perforating vessels: High signal-to-noise ratio signals located adjacent to each other may be subcutaneous veins in addition to the target perforating vessels. Both have high signal intensity in the above process, but the main difference between them is that perforating vessels continue to advance deep and eventually come into contact with the muscle surface, while subcutaneous veins advance parallel to the fat layer throughout and do not come into contact with deep tissue (see Figure 7, where C is the signal of a perforating vessel and D is the signal of a vein). Therefore, in this process, venous signals are excluded by distinguishing between subcutaneous veins and perforating vessels using the logic for determining the slope of the signal relative to the deep muscle layer and the logic for determining the signal depth obtained in the previous process.

[0036] 10) After marking the perforating vessels, score them comprehensively to obtain the optimal perforating vessel. After eliminating venous signals, multiple perforating vessels may be present. To ensure the success rate of the surgery, in this embodiment, each perforating vessel is scored, and the optimal perforating vessel is made the final perforating vessel. When scoring in this step, multiple perforating vessels are comprehensively scored using a combination of the signal-to-noise ratio of the perforating vessels and logic favorable to the determination of the quality of other confirmed vessels (e.g., outer diameter, branching). The scoring may also be performed using a weighted sum method. After calculating the score, the location of the candidate perforating vessel with the highest score is marked, and this location becomes the identified optimal perforating vessel.

[0037] 11) The coordinates of the muscle penetration points of the optimal perforating vessels are converted to a body surface coordinate system. Specifically, the spatial position information of the muscle penetration points of the perforating vessels is projected onto the body surface, and a coordinate system is constructed with an arbitrary mark point within the target area of ​​the skin on the body surface as the origin. The algorithm then calculates the coordinates of the projected point in the body surface coordinate system, making it easier for the user to mark the position information on the patient's body surface, thereby providing support for intraoperative design.

[0038] The perforating vessels obtained by the above method basically meet the gold standard, with an identification rate of over 99%. The surgeon can automatically review the images of the identified perforating vessels, confirm the results, and select the candidate perforating vessel that best matches the requirements of the surgical design. Here, the gold standard is the result of artificial judgment by a physician who is well-versed in the medical imaging technology of tomography for identifying perforating vessels.

[0039] A preliminary test was conducted using data from 20 actual cases (12 cases involving subabdominal perforating vessels and 8 cases involving perforating vessels of the descending lateral femoral artery). The calculation results of this embodiment were compared with the average results of three clinicians with more than five years of experience in artificial identification of perforating vessels. The agreement rate for the top three perforating vessels was 98.33%, for the top five perforating vessels it was 96%, and for the top ten perforating vessels it was 90.50% (see Figure 8).

[0040] As is readily apparent, this invention reduces 3D image data to 2D through dimensionality reduction processing, and then searches for the signal points of perforating vessels in the 1D image. By improving processing speed, the task of locating the perforating vessels can be completed within 5 minutes, significantly reducing the time required for locating perforating vessels compared to the original 45 minutes, thereby improving work efficiency.

[0041] A second embodiment of the present invention relates to a computer system comprising a storage device and a processor, wherein a computer program is stored in the storage device, and when the computer program is executed by the processor, the processor performs the steps of the method for automatically identifying and locating perforating vessels in the first embodiment. This embodiment can be mounted on many mobile computing platforms to enable mobile or remote operation of the identification work (see Figure 9).

[0042] A third embodiment of the present invention relates to a readable storage medium in which an executable computer program is stored, and when the executable computer program is executed by a processor, the processor performs the steps of the method for automatically identifying and locating perforating vessels in the first embodiment.

Claims

1. A method for automatically identifying and locating perforating vessels, The process involves determining the scanning range in the scanning direction from the target perforating branch, and then extracting three-dimensional image data within the scanning range. The process of obtaining layered image data by performing dimensionality reduction processing on the aforementioned three-dimensional image data, The process involves using the aforementioned stratified image data as the minimum analysis unit and screening signal points of perforating vessels using the signal-to-noise ratio dynamic threshold method, The process involves three-dimensional reconstruction of the signal points of the connected perforating vessels in the three-dimensional image data, and screening candidate perforating vessels using the total signal-to-noise ratio score after reconstruction. A process to identify perforating vessels based on their relative position, inclination, and depth relative to the epidermal and deep muscular layers, Includes, The process of obtaining layered image data by performing dimensionality reduction processing on the aforementioned three-dimensional image data is, specifically, The method involves merging n consecutive images in 3D image data to obtain signal-enhanced 2D image data, In two-dimensional image data, the boundaries of the skin surface and the boundaries between subcutaneous fat and muscle are marked, and the subcutaneous fat region enclosed by these two boundaries is designated as the target analysis area. The analysis region of the target is stratified to obtain multiple consecutive stratified image data of subcutaneous depth, Includes, A method for automatically identifying and locating perforating vessels, characterized in that when merging n consecutive images in the aforementioned three-dimensional image data, the point with the smallest pixel value at the same location is used as the pixel value of the merged image.

2. A method for automatically identifying and locating perforating vessels, The process involves determining the scanning range in the scanning direction from the target perforating branch, and then extracting three-dimensional image data within the scanning range. The process of obtaining layered image data by performing dimensionality reduction processing on the aforementioned three-dimensional image data, The process involves using the aforementioned stratified image data as the minimum analysis unit and screening signal points of perforating vessels using the signal-to-noise ratio dynamic threshold method, The process involves three-dimensional reconstruction of the signal points of the connected perforating vessels in the three-dimensional image data, and screening candidate perforating vessels using the total signal-to-noise ratio score after reconstruction. A process to identify perforating vessels based on their relative position, inclination, and depth relative to the epidermal and deep muscular layers, Includes, The process of obtaining layered image data by performing dimensionality reduction processing on the aforementioned three-dimensional image data is, specifically, The method involves merging n consecutive images in 3D image data to obtain signal-enhanced 2D image data, In two-dimensional image data, the boundaries of the skin surface and the boundaries between subcutaneous fat and muscle are marked, and the subcutaneous fat region enclosed by these two boundaries is designated as the target analysis area. The analysis region of the target is stratified to obtain multiple consecutive stratified image data of subcutaneous depth, Includes, A method for automatically identifying and locating perforating vessels, characterized in that the boundary coordinates between the air and the skin surface are obtained by a machine learning algorithm, and the boundary coordinates between the subcutaneous fat and muscle are obtained from the rate of change of the average pixel value.

3. A method for automatically identifying and locating perforating vessels, The process involves determining the scanning range in the scanning direction from the target perforating branch, and then extracting three-dimensional image data within the scanning range. The process of obtaining layered image data by performing dimensionality reduction processing on the aforementioned three-dimensional image data, The process involves using the aforementioned stratified image data as the minimum analysis unit and screening signal points of perforating vessels using the signal-to-noise ratio dynamic threshold method, The process involves three-dimensional reconstruction of the signal points of the connected perforating vessels in the three-dimensional image data, and screening candidate perforating vessels using the total signal-to-noise ratio score after reconstruction. A process to identify perforating vessels based on their relative position, inclination, and depth relative to the epidermal and deep muscular layers, Includes, The process of screening signal points of perforating vessels using the aforementioned stratified image data as the minimum analysis unit and the signal-to-noise ratio dynamic threshold method is, specifically, The same subcutaneous depth stratified image data is converted into a two-dimensional array, the minimum or maximum pixel value at several pixel points at different depths in the two-dimensional array at similar skin locations is designated as a candidate vascular signal point, and the minimum or maximum pixel value at a nearby location is designated as the background reference value, and the relative signal-to-noise ratio of the candidate vascular signal points is calculated. A set of points is formed by selecting points whose pixel values ​​are smaller than the pixel threshold, and from this set, points where the relative signal-to-noise ratio exceeds the signal-to-noise ratio threshold are selected as signal points for perforating vessels, and the horizontal coordinates corresponding to the signal points for the perforating vessels are marked. A method for automatically identifying and locating perforating vessels, characterized by including the following:

4. When calculating the relative signal-to-noise ratio of the candidate vascular signal points, ((X ４ + X ５ / 2) * (X ２ + X ３ + X ４ + X ５ )) / (X ０ + X １ ), where X ｎ is the average value of the pixels n pixels away from the candidate vascular signal point, and X ０ is the pixel value of the candidate vascular signal point. A method for automatically identifying and localizing the perforating branch vessels according to claim 3, characterized in that.

5. A method for automatically identifying and locating perforating vessels, The process involves determining the scanning range in the scanning direction from the target perforating branch, and then extracting three-dimensional image data within the scanning range. The process of obtaining layered image data by performing dimensionality reduction processing on the aforementioned three-dimensional image data, The process involves using the aforementioned stratified image data as the minimum analysis unit and screening signal points of perforating vessels using the signal-to-noise ratio dynamic threshold method, The process involves three-dimensional reconstruction of the signal points of the connected perforating vessels in the three-dimensional image data, and screening candidate perforating vessels using the total signal-to-noise ratio score after reconstruction. A process to identify perforating vessels based on their relative position, inclination, and depth relative to the epidermal and deep muscular layers, Includes, The process of three-dimensionally reconstructing the signal points of the connected perforating vessels in the aforementioned three-dimensional image data is, specifically, The process involves projecting the transverse coordinates corresponding to the signal points of the perforating vessels onto two-dimensional image data, merging points adjacent to the signal points of the perforating vessels in the two-dimensional image data, and calculating the sum to determine the continuity of the signal points of the merged perforating vessels, the spatial propagation method, and the signal-to-noise ratio. The process involves fusing two-dimensional image data to form a three-dimensional space, placing the merged points from the two-dimensional image data into the three-dimensional space, merging them with adjacent points in the three-dimensional space, and then calculating the sum to determine the signal-to-noise ratio of the merged data. A method for automatically identifying and locating perforating vessels, characterized by including the following:

6. A method for automatically identifying and locating perforating vessels according to any one of claims 1 to 5, further comprising the step of calculating a total score for each perforating vessel based on the signal-to-noise ratio and vascular mass after the above step (5), and designating the perforating vessel with the highest total score as the final perforating vessel.

7. A computer device comprising a storage device and a processor, wherein a computer program is stored in the storage device, and when the computer program is executed by the processor, the processor performs the steps of a method for automatically identifying and locating perforating vessels as described in any one of claims 1 to 5.

8. A computer-readable storage medium wherein an executable computer program is stored in the computer-readable storage medium, and when the executable computer program is executed by a processor, the processor performs the steps of a method for automatically identifying and locating perforating vessels as described in any one of claims 1 to 5.

Images