A fully automated method for detecting ASIS in abdominal and pelvic CT volumes
Through steps such as cropping and threshold setting, automated detection of ASIS positions in CT images is realized, solving complex operation problems in the prior art, and improving the efficiency and accuracy of medical imaging operations.
Patent Information
- Application Number
- CN202111247344.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-10-26
- Publication Date
- 2024-11-01
- Estimated Expiration
- 2041-10-26
AI Technical Summary
The operation method of detecting ASIS positions in CT images in the prior art is complex, which affects the working efficiency of medical imaging operations and lacks automated measurement methods.
By cropping the full-body CT image, setting the threshold interval, establishing a coordinate system, conducting horizontal slice searches, setting the L_ROI and R_ROI areas, confirming the ASIS candidate positions step by step, and realizing automated detection.
It improves image processing efficiency and accuracy, realizes automated detection of ASIS locations, and simplifies medical imaging operation process.
Smart Images

Figure CN113989370B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of medical imaging, and in particular to a method for fully automatic detection of ASIS in an abdominal and pelvic CT volume. Background Art
[0002] The ASIS (anterior superior iliac spine) is a bone structure in the human anterior pelvis that is easily detected (by simple palpation) and is often used as a reference landmark for different medical applications. However, the existing technology usually uses manual resolution to find the location of the ASIS in CT images, which is relatively complicated and seriously affects the efficiency of medical imaging operations. Therefore, it is necessary to study a method for automatically measuring the location of the ASIS in CT images to solve the above problem. Summary of the invention
[0003] The object of the present invention is to provide a method for fully automatic detection of ASIS in the abdominal and pelvic CT volume with improved operation efficiency in view of the above problems.
[0004] In order to achieve the above object, the technical solution of the present invention is:
[0005] A fully automated method for detecting ASIS within an abdominal and pelvic CT volume comprises the following steps:
[0006] S1, cropping the part above the mid-axis plane in the whole-body CT image to obtain a lower body image;
[0007] S2, set the lower threshold LT and the upper threshold UT, set the voxel values in the interval [LT, UT] in the lower body image as foreground pixels, and set the voxel values outside the interval [LT, UT] in the lower body image as background pixels;
[0008] S3, establishing a coordinate system and placing the lower body image in the coordinate system;
[0009] S4, horizontally slicing the lower body image from top to bottom, and searching the slices starting from the topmost slice in the lower body image;
[0010] S5. Find the L_ASIS candidate position and the R_ASIS candidate position in the current slice according to the foreground pixels and the background pixels;
[0011] S6. If the i - coordinate values of the L_ASIS candidate positions and R_ASIS candidate positions in the current slice are lower than those of the L_ASIS candidate positions and R_ASIS candidate positions in the upper - level slice, accept the L_ASIS candidate positions and R_ASIS candidate positions in the current slice and select the next - level slice to continue repeating step S6; if the i - coordinate values of the L_ASIS candidate positions and R_ASIS candidate positions in the current slice are not lower than those of the L_ASIS candidate positions and R_ASIS candidate positions in the upper - level slice, select to end.
[0012] S7. Finally, it is determined that the L_ASIS candidate positions and R_ASIS candidate positions with the lowest i - coordinate values in the slice are the final detection positions.
[0013] Further, in step S2, the lower threshold LT = WL - WW / 2, and the upper threshold UT = WL + WW / 2; where WL and WW are statistical estimation constants that only retain bone tissue.
[0014] Further, in step S5, finding the L_ASIS candidate positions and R_ASIS candidate positions in the current slice includes the following steps:
[0015] S51. Set the L_ROI region and R_ROI region in the slice, search for the L_ASIS candidate positions in the L_ROI region, and search for the R_ASIS candidate positions in the R_ROI region.
[0016] S52. In the slice, divide the part with coordinate value j > Jcutoff as the L_ROI region.
[0017] S53. In the slice, divide the part with coordinate value j < N - Jcutoff as the R_ROI region.
[0018] S54. Find the foreground pixels in the L_ROI region, then the L_ASIS candidate position is (i0, j0), where i0 ≤ all i - coordinate values in the L_ROI region.
[0019] S55. Find the foreground pixels in the R_ROI region, then the R_ASIS candidate position is (i1, j1), where i1 ≤ all i - coordinate values in the R_ROI region.
[0020] Further, the Jcutoff is 0.25 * N.
[0021] Compared with the prior art, the advantages and positive effects of the present invention are:
[0022] The present invention firstly reduces the amount of data processing by cropping the whole image of the human body, thereby facilitating the processing and analysis of the human body image and improving the image processing efficiency; then removes the confusion caused by the spine by setting the L_ROI area and the R_ROI area, thereby improving the processing accuracy of the image; finally, the position of the ASIS is confirmed by searching for the L_ASIS candidate position and the R_ASIS candidate position in each level of slice; it realizes the effect of automatically detecting the ASIS position, improves the work efficiency of medical imaging operations, and brings convenience to medical imaging work. BRIEF DESCRIPTION OF THE DRAWINGS
[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 or the description of the prior art 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 labor.
[0024] Figure 1 It is a schematic diagram of the structure of the coordinate system;
[0025] Figure 2 This is a schematic diagram of the location of ASIS;
[0026] Figure 3 It is a schematic diagram of the structure of ASIS in a slice;
[0027] Figure 4 A diagram showing the position relationship of multiple slices in the coordinate system. DETAILED DESCRIPTION
[0028] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work, any modifications, equivalent substitutions, improvements, etc., should be included in the protection scope of the present invention.
[0029] like Figures 1 to 4 As shown, this embodiment discloses a fully automatic method for detecting ASIS in the abdominal and pelvic CT volume, where ASIS includes left ASIS and right ASIS. The search for left ASIS and right ASIS is performed slice by slice from top to bottom in a stack of axial CT slices that conform to the volume scan (size is MxNxH).
[0030] The process is mainly divided into two stages:
[0031] 1. Preprocessing
[0032] 1.1. Cropping
[0033] Since both ASIS are located in the pelvis, which is always in the lower half of the scan, the upper half is cropped away, resulting in a lower body image. This helps increase throughput while eliminating rib confusion (due to less processing).
[0034] 1.2 Binarization
[0035] The lower threshold (LT) and the upper threshold (UT) are set, and the voxel values in the range [LT, UT] in the lower body image are set as foreground pixels (1), otherwise they are set as background pixels (0).
[0036] LT=WL-WW / 2, UT=WL+WW / 2;
[0037] Where WL and WW are statistical estimation constants that retain only bone tissue.
[0038] 2. Search
[0039] 2.1. Establish a coordinate system and place the lower body image in the coordinate system; in the coordinate system, the width direction of the human body (i.e., the left-right direction of the human body) is used as the X-axis, the thickness direction of the human body (i.e., the front-back direction of the human body) is used as the Y-axis, and the height of the human body (i.e., the up-down direction of the human body) is used as the Z-axis;
[0040] 2.2. Slice the lower body image horizontally from top to bottom, and search the slices one by one starting from the top slice in the lower body image, and start from slice k=H / 2, where H is the height of the person;
[0041] 2.3. Assuming that the left and right ASIS can be (approximately) found in the same slice, this is a good approximation for detection purposes.
[0042] 2.4. Find the L_ASIS candidate position and the R_ASIS candidate position in the current slice based on the foreground pixels and the background pixels. The specific operations are as follows:
[0043] 2.4.1. In order to eliminate the confusion of the part close to the spine in each slice, the L_ROI region and the R_ROI region are set in the slice, and the L_ASIS candidate position is searched in the L_ROI region, and the R_ASIS candidate position is searched in the R_ROI region;
[0044] 2.4.2. In the slice, the portion with coordinate value j>Jcutoff is divided as L_ROI area; Jcutoff is 0.25*N;
[0045] 2.4.3. Divide the part with the coordinate value j < N - Jcutoff in the slice as the R_ROI region; Jcutoff is 0.25 * N; the coordinate value j represents the sequence number of the graph, such as Figure 1 as shown;
[0046] 2.4.4. Find the foreground pixel (1) in the L_ROI region, then the candidate position of L_ASIS is (i0, j0), where i0 ≤ all the i coordinate values in the L_ROI region; (i0, j0) represents an initialized search coordinate position;
[0047] 2.4.5. Find the foreground pixel (1) in the R_ROI region, then the candidate position of R_ASIS is (i1, j1), where i1 ≤ all the i coordinate values in the R_ROI region; (i1, j1) represents the updated coordinate position after search;
[0048] 2.5. If the i coordinate values of the candidate positions of L_ASIS and R_ASIS in the current slice are lower than those of the candidate positions of L_ASIS and R_ASIS in the previous slice, then accept the candidate positions of L_ASIS and R_ASIS in the current slice and select the next slice to continue repeating step S6; if the i coordinate values of the candidate positions of L_ASIS and R_ASIS in the current slice are not lower than those of the candidate positions of L_ASIS and R_ASIS in the previous slice, then select to end; the i coordinate value represents the horizontal row of the graph, such as Figure 1 as shown;
[0049] 2.6. Finally, it is determined that the candidate positions of L_ASIS and R_ASIS with the lowest i coordinate values in the slice are the final detection positions.
[0050] First, the present invention reduces the data processing amount by cropping the overall image of the human body, thus facilitating the processing and analysis of the human body image and improving the image processing efficiency; then, by setting the L_ROI region and the R_ROI region to remove the confusion caused by the spine, the processing accuracy of the image is improved; finally, by searching for the candidate positions of L_ASIS and R_ASIS in each slice to confirm the positions of ASIS; it realizes the effect of automatically detecting the positions of ASIS, improves the working efficiency of medical image operations, and brings convenience to medical image work.
Claims
1. A fully automated method for detecting ASIS in an abdominal and pelvic CT volume, characterized by: It includes the following steps: S1. Crop the part above the mid-axis plane in the whole-body CT image to obtain the lower-body image; S2. Set the lower threshold LT and the upper threshold UT, set the voxel values within the interval [LT, UT] in the lower-body image as foreground pixels, and set the voxel values outside the interval [LT, UT] in the lower-body image as background pixels; S3. Establish a coordinate system and place the lower-body image in the coordinate system; S4. Horizontally slice the lower-body image from top to bottom, and start searching for the slices from the topmost slice in the lower-body image; S5. According to the foreground pixels and background pixels, find the L_ASIS candidate positions and R_ASIS candidate positions in the current slice; S6. If the i coordinate values of the L_ASIS candidate positions and R_ASIS candidate positions in the current slice are lower than the i coordinate values of the L_ASIS candidate positions and R_ASIS candidate positions in the upper-level slice, accept the L_ASIS candidate positions and R_ASIS candidate positions in the current slice and select the next-level slice to continue repeating step S6; if the i coordinate values of the L_ASIS candidate positions and R_ASIS candidate positions in the current slice are not lower than the i coordinate values of the L_ASIS candidate positions and R_ASIS candidate positions in the upper-level slice, then select to end; S7. Finally determine that the L_ASIS candidate position and R_ASIS candidate position with the lowest i coordinate value in the slice are the final detection positions; In step S5, finding the L_ASIS candidate positions and R_ASIS candidate positions in the current slice includes the following steps: S51. Set the L_ROI region and R_ROI region in the slice, search for the L_ASIS candidate position in the L_ROI region, and search for the R_ASIS candidate position in the R_ROI region; S52. Divide the part with the coordinate value j > Jcutoff in the slice as the L_ROI region; S53. Divide the part with the coordinate value j < N - Jcutoff in the slice as the R_ROI region; S54. Find the foreground pixels in the L_ROI region, then the L_ASIS candidate position is (i0, j0), where i0 ≤ all the i coordinate values in the L_ROI region; S55. Find the foreground pixels in the R_ROI region, then the R_ASIS candidate position is (i1, j1), where i1 ≤ all the i coordinate values in the R_ROI region.
2. The method of fully automatic detection of ASIS in an abdominal and pelvic CT volume as claimed in claim 1, characterized in that: In step S2, the lower threshold LT = WL - WW / 2, and the upper threshold UT = WL + WW / 2; where WL and WW are statistical estimation constants for only retaining bone tissue.
3. The method of fully automatic detection of ASIS in an abdominal and pelvic CT volume as claimed in claim 1, characterized in that: The Jcutoff is 0.25 * N.
Citation Information
Patent Citations
Method for establishing pelvis CT three-dimensional coordinate system
CN106780715A
Front human body image segmentation method and a device under a complex background
CN109509194A
Ultrasonic image and CT image registration method, device and equipment
CN112184781A
Automatic Removal of Physiological Accumulation from Nuclear Medicine Image, and Automatic Segmentation of CT Image
US20190150860A1
