A CT image rapid registration method and system
By calculating lesion characteristics and artifact characteristics within the contour area of the lung tissue and screening key pixels for image registration, the problem of accuracy of CT and CBCT images was solved, and the accuracy of radiotherapy plans was improved.
Patent Information
- Application Number
- CN202411073896.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-07
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2044-08-07
AI Technical Summary
The prior art cannot effectively ensure the accurate registration of CT images and CBCT images, resulting in inaccurate radiotherapy planning, mainly due to artifacts and position differences caused by grayscale differences, structural inconsistencies and physiological movements of human bodies.
By obtaining the degree of lesions of pixel points in the contour area of the lung tissue, the suspected lesions are screened out, the degree of artifact characteristics is calculated, artifact interference is eliminated, the key lesions pixel points are obtained, and image registration is performed based on these points.
It improves the registration accuracy of CT images and CBCT images, reduces the interference of artifacts and organ movement on lesion area identification, and ensures the accuracy of radiotherapy planning.
Smart Images

Figure CN118587263B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of image processing, and in particular to a method and system for rapid CT image registration. Background Art
[0002] CBCT technology is an image-guided technology widely used in radiotherapy. It can monitor the status of tumors and normal organs. Radiotherapy doctors usually align the positioning CT image and the CBCT image to achieve accurate radiation of the tumor. Due to the grayscale difference and structural inconsistency between CT and CBCT images, the physiological movement of the human body causes artifacts in the CBCT image, and the difference in positioning of each shot causes the imaging position of the lesion to shift. Therefore, the existing technology cannot effectively guarantee the accuracy of the alignment of the positioning CT image and the CBCT image, which makes the radiotherapy plan inaccurate. Summary of the invention
[0003] In order to solve the above problems, the present invention provides a CT image rapid registration method and system.
[0004] An embodiment of the present invention provides a method for rapid CT image registration, the method comprising the following steps:
[0005] Acquire a positioning CT image and several CBCT images of the patient;
[0006] According to the grayscale difference between each pixel point and the surrounding pixels in the lung tissue contour area in the CBCT image, the degree of conformity with the lesion characteristics of each pixel point in the lung tissue contour area is obtained; based on the degree of conformity with the lesion characteristics, all the pixels in the lung tissue contour area are screened to obtain the suspected lesion area; according to the degree of conformity with the lesion characteristics and the position distribution of the edge pixels in the suspected lesion area, the possibility of the suspected lesion area belonging to the lesion area is obtained; based on the possibility of belonging to the lesion area, all suspected lesion areas are screened to obtain the suspected target lesion area;
[0007] According to the suspected target lesion areas in different CBCT images, all suspected target lesion pixel points are obtained; according to the difference between the degree of compliance with lesion characteristics and the possibility of belonging to the lesion area of the suspected target lesion pixel points in different CBCT images, the degree of compliance with artifact characteristics of each suspected target lesion pixel point is obtained; all lesion areas in the CBCT image are obtained based on the degree of compliance with artifact characteristics; according to the difference in the degree of compliance with artifact characteristics between each edge pixel point and the adjacent edge pixel points in the lesion area, all key lesion pixel points in the CBCT image are obtained;
[0008] According to the difference in the degree of conformity of the key lesion pixel points with all the pixels in the positioning CT image, the matching pixel points of each key lesion pixel point in the positioning CT image are obtained; based on the matching pixel points of the key lesion pixel points in the positioning CT image, the CBCT image and the positioning CT image are registered.
[0009] Preferably, the method of obtaining the degree of conformity with the lesion characteristics of each pixel in the lung tissue contour area according to the grayscale difference between each pixel and the surrounding pixels in the lung tissue contour area in the CBCT image includes the following specific methods:
[0010] For any pixel point in the lung tissue contour area in any CBCT image, the grayscale difference between the pixel point and the pixels in its 8 neighborhoods is recorded as the grayscale change degree in the neighborhood of the pixel point; the product of the mean of the grayscale values of all pixels in the lung tissue contour area in the CBCT image and the grayscale change degree in the neighborhood of the pixel point is recorded as the first product; the normalized value of the ratio between the first product and the mean of the grayscale values of all pixels in the 8 neighborhoods of the pixel point is taken as the degree of compliance with the lesion characteristics of the pixel point.
[0011] Preferably, the method of screening all pixel points in the lung tissue contour area based on the degree of conformity with the lesion characteristics to obtain the suspected lesion area includes the following specific methods:
[0012] Preset a lesion threshold parameter For any pixel point in the lung tissue contour area in any CBCT image, if the degree of conformity of the pixel point to the lesion feature is greater than or equal to the lesion threshold parameter , the pixel point is recorded as a suspected lesion pixel point; and multiple regions formed by all the suspected lesion pixel points in the CBCT image are recorded as suspected lesion regions.
[0013] Preferably, the possibility of the suspected lesion area belonging to the lesion area is obtained according to the degree of compliance with the lesion characteristics and the position distribution of the edge pixels in the suspected lesion area, including the specific method of:
[0014] For any CBCT image in the lung tissue contour area If there are two edge pixels on the same horizontal line in the suspected lesion area, and the two edge pixels are located on both sides of the central axis of the CBCT image, the two edge pixels are recorded as valid pixel points of the suspected lesion area, and record the combination of the two edge pixel points as a valid pixel point combination;
[0015] For For any valid pixel point in any valid pixel point combination of the suspected lesion area, the Euclidean distance between the valid pixel point and the central axis of the CBCT image is recorded as the first distance of the valid pixel point; the absolute value of the difference between the first distances between two valid pixel points in the valid pixel point combination is recorded as the lesion characteristic distance of the valid pixel point combination;
[0016] The first The ratio between the mean value of the gradient amplitude of all edge pixels in the suspected lesion area and the mean value of the gradient amplitude of all edge pixels in all suspected lesion areas in the lung tissue contour area in the CBCT image is recorded as the first ratio; The ratio between the mean of the degree of conformity to lesion characteristics of all edge pixels of the first suspected lesion area and the mean of the degree of conformity to lesion characteristics of all edge pixels of all suspected lesion areas in the lung tissue contour area in the CBCT image is recorded as the second ratio; The number of all valid pixels in the suspected lesion area is The ratio of the number of all edge pixels in the suspected lesion area is recorded as the third ratio; The cumulative sum of the lesion feature distances of the effective pixel point combinations of the suspected lesion area is recorded as the first cumulative sum; the normalized value of the product of the first ratio, the second ratio, the third ratio and the first cumulative sum is recorded as the second cumulative sum. The possibility that the suspected lesion area belongs to the lesion area.
[0017] Preferably, the specific method of acquiring all suspected target lesion pixels according to the suspected target lesion areas in different CBCT images includes:
[0018] For any CBCT image, the grayscale values of all pixels in all suspected target lesion areas in the CBCT image are set to 1, and the other pixels are set to 0 to obtain a binary image of the CBCT image; the binary images of all CBCT images are subjected to a difference operation to obtain a CBCT difference image; for any pixel in the CBCT difference image, if the grayscale value of the pixel is not 0, the pixel is recorded as a suspected target lesion pixel.
[0019] Preferably, the method of obtaining the degree of conformity to artifact features of each suspected target lesion pixel point according to the difference between the degree of conformity to lesion features and the possibility of belonging to the lesion area of the suspected target lesion pixel point in different CBCT images includes the following specific methods:
[0020] For any suspected target lesion pixel point, the ratio between the mean of the degree of conformity to lesion characteristics of all suspected target lesion pixels in all CBCT images and the mean of the degree of conformity to lesion characteristics of the suspected target lesion pixels in all CBCT images is recorded as the suspected lesion ratio; the ratio between the mean of the probability of belonging to the lesion area of the suspected lesion area to which all suspected target lesion pixels belong in all CBCT images and the mean of the probability of belonging to the lesion area of the suspected lesion area to which the suspected target lesion pixels belong in all CBCT images is recorded as the lesion ratio;
[0021] The difference between the degree of conformity of the suspected target lesion pixel point with the lesion feature in different CBCT images is used to obtain the degree of change of the lesion feature of the suspected target lesion pixel point;
[0022] The normalized value of the product of the lesion feature change degree of the suspected target lesion pixel, the lesion ratio and the suspected lesion ratio is used as the conforming artifact feature degree of the suspected target lesion pixel.
[0023] Preferably, the method of acquiring all lesion areas in the CBCT image based on the degree of conformity with the artifact characteristics includes:
[0024] Jordi If the degree of conformity of the artifact characteristics of the suspected target lesion pixel point is greater than or equal to the preset threshold parameter, the first The suspected target lesion pixel points are recorded as lesion pixels; the multiple regions formed by all lesion pixels in the CBCT difference image correspond to the multiple regions in the CBCT image, which are all recorded as lesion regions.
[0025] Preferably, the method of obtaining all key lesion pixels in the CBCT image according to the difference in the degree of conformity to the artifact characteristics between each edge pixel in the lesion area and the adjacent edge pixel points includes the following specific methods:
[0026] In the lesion area The average of the degree of conformity of all pixels in the 8-neighborhood of the edge pixel is The absolute value of the difference between the degree of conformity to the artifact feature of the edge pixels is recorded as the first absolute value of the difference; The tangent vector of the edge pixel is The cosine value of the angle between the tangent vectors of edge pixels is recorded as The turning degree of the edge pixel point; the absolute value of the first difference and the The normalized value of the product of the turning degrees of the edge pixels is recorded as The probability of critical lesions at edge pixels;
[0027] Preset a key threshold parameter , The probability of critical lesions of edge pixels is greater than or equal to the critical threshold parameter , will The edge pixels are recorded as key lesion pixels.
[0028] Preferably, the method of obtaining the matching pixel point of each key lesion pixel point in the positioning CT image according to the difference in the degree of conformity with the lesion characteristics between the key lesion pixel point and all the pixels in the positioning CT image includes the following specific methods:
[0029] For any pixel point in the localized CT image, The absolute value of the difference between the degree of conformity to the lesion characteristics of the key lesion pixel point and the degree of conformity to the lesion characteristics of the pixel point is recorded as the target difference; The Euclidean distance between the key lesion pixel and the pixel in the image is recorded as the target distance; the reciprocal of the product of the target difference and the target distance is recorded as The matching coefficient between a key lesion pixel point and the pixel point;
[0030] In the Among the matching coefficients between the key lesion pixel points and all the pixels in the positioning CT image, the pixel with the largest matching coefficient is recorded as The matching pixels of the key lesion pixels.
[0031] The present invention also proposes a CT image rapid registration system, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements any one of the steps of the CT image rapid registration method when executing the computer program.
[0032] The technical solution of the present invention has the following beneficial effects: the present invention obtains the degree of conformity to artifact characteristics of each suspected target lesion pixel point according to the difference between the degree of conformity to lesion characteristics and the possibility of belonging to the lesion area of the suspected target lesion pixel point in different CBCT images; obtains all lesion areas in the CBCT image based on the degree of conformity to artifact characteristics; thereby eliminating the interference of artifacts and organ movement on the identification of lesion areas, thereby obtaining accurate lesion areas; according to the difference in the degree of conformity to artifact characteristics between each edge pixel point and adjacent edge pixels in the lesion area, obtains all key lesion pixels in the CBCT image; according to the difference in the degree of conformity to lesion characteristics between the key lesion pixel point and all pixels in the positioning CT image, obtains the matching pixel point of each key lesion pixel point in the positioning CT image; based on the matching pixel point of the key lesion pixel point in the positioning CT image, the CBCT image is registered with the positioning CT image; thereby, the CBCT image is registered with the positioning CT image based on the key lesion pixel point with higher recognizability, thereby improving the accuracy of registration between the positioning CT image and the CBCT image. BRIEF DESCRIPTION OF THE DRAWINGS
[0033] 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 work.
[0034] Figure 1 A flowchart of the steps of a method for rapid CT image registration according to the present invention;
[0035] Figure 2 This is a feature relationship flow chart of a rapid CT image registration method according to the present invention. DETAILED DESCRIPTION
[0036] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of a CT image rapid registration method and system proposed according to the present invention, its specific implementation, structure, features and effects, in conjunction with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.
[0037] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
[0038] The specific scheme of a CT image rapid registration method and system provided by the present invention is described in detail below with reference to the accompanying drawings.
[0039] See also Figure 1 , which shows a flowchart of a method for rapid CT image registration provided by an embodiment of the present invention, the method comprising the following steps:
[0040] Step S001: Acquire a positioning CT image and several CBCT images of the patient.
[0041] It should be noted that, taking the radiotherapy process of lung cancer as an example, before radiotherapy, the doctor will formulate a radiotherapy plan based on the positioning CT. However, since the position of human organs may change every day, and in actual operation it is impossible to guarantee that the position of the human body is exactly the same as the position in the positioning CT, it is usually taken once a day before radiotherapy to verify the CBCT (cone beam CT) image to ensure the accuracy of treatment.
[0042] Specifically, we first need to collect several CBCT images of the patient. The specific process is as follows:
[0043] A positioning CT image is obtained; within three days before the patient undergoes radiotherapy, a CBCT image of the patient is obtained by a linear accelerator every day; and then several CBCT images of the patient are obtained.
[0044] So far, the positioning CT image and several CBCT images of the patient are obtained through the above method.
[0045] Step S002: According to the grayscale difference between each pixel point and the surrounding pixels in the lung tissue contour area in the CBCT image, the degree of compliance with the lesion characteristics of each pixel point in the lung tissue contour area is obtained; based on the degree of compliance with the lesion characteristics, all pixels in the lung tissue contour area are screened to obtain a suspected lesion area; according to the degree of compliance with the lesion characteristics and the position distribution of the edge pixels in the suspected lesion area, the possibility of the suspected lesion area belonging to the lesion area is obtained; based on the possibility of belonging to the lesion area, all suspected lesion areas are screened to obtain a suspected target lesion area.
[0046] 1. Obtain all suspected lesion areas in each CBCT image.
[0047] It should be noted that since the lesion area has a weaker X-ray absorption capacity than the normal human body structure and the density of the lesion area is generally uneven, it is displayed as an uneven low-density shadow in the CBCT image. Based on this feature, the degree of conformity with the lesion characteristics of each pixel point in the lung tissue contour area is obtained.
[0048] Preferably, for any CBCT image, the lung tissue contour area in the CBCT image is obtained using the Otsu threshold segmentation algorithm;
[0049] It should be noted that the Otsu threshold segmentation algorithm is a well-known technical means well known to those skilled in the art, and will not be described in detail in this embodiment.
[0050] Preferably, in some implementations of the embodiments of the present invention, according to the grayscale difference between each pixel point and the surrounding pixels in the lung tissue contour area in the CBCT image, the specific method for obtaining the degree of compliance with the lesion characteristics of each pixel point in the lung tissue contour area is:
[0051] For any pixel point in the lung tissue contour area in any CBCT image, the grayscale difference between the pixel point and the pixels in its 8 neighborhoods is recorded as the grayscale change degree in the neighborhood of the pixel point; the product of the mean of the grayscale values of all pixels in the lung tissue contour area in the CBCT image and the grayscale change degree in the neighborhood of the pixel point is recorded as the first product; the normalized value of the ratio between the first product and the mean of the grayscale values of all pixels in the 8 neighborhoods of the pixel point is taken as the degree of compliance with the lesion characteristics of the pixel point.
[0052] The specific formula is:
[0053]
[0054] In the formula, Indicates the first The degree to which each pixel point conforms to the lesion characteristics; It represents the mean gray value of all pixels in the lung tissue contour area in the CBCT image; Indicates The number of all pixels in the 8-neighborhood of a pixel; Indicates The 8-neighborhood of the pixel The gray value of each pixel; Indicates The gray value of each pixel; Indicates The mean gray value of all pixels in the 8-neighborhood of a pixel; represents the linear normalization function.
[0055] It should be noted that Indicates The grayscale change degree in the neighborhood of a pixel point. The larger the value, the The more uneven the grayscale distribution in the neighborhood of a pixel point is; Indicates The gray value in the neighborhood of a pixel point. The larger the value, the The more pixels there are, the more they conform to the characteristics of the lesion.
[0056] Preferably, in some implementations of the embodiments of the present invention, all pixels in the lung tissue contour area are screened based on the degree of conformity with the lesion characteristics, and the specific method for obtaining the suspected lesion area is:
[0057] Preset a lesion threshold parameter , wherein this embodiment is based on This example is described as an example, and this embodiment is not specifically limited. Depends on the specific implementation situation;
[0058] For any pixel point in the lung tissue contour area in any CBCT image, if the degree of conformity of the pixel point to the lesion feature is greater than or equal to the lesion threshold parameter , the pixel point is recorded as a suspected lesion pixel point; and multiple regions formed by all the suspected lesion pixel points in the CBCT image are recorded as suspected lesion regions.
[0059] At this point, all suspected lesion areas in each CBCT image are obtained.
[0060] 2. Obtain all suspected target lesion areas in each CBCT image.
[0061] It should be noted that since the edge of the lesion area is not smooth and the grayscale difference between it and the surrounding area is not as great as that between the normal organ and the surrounding area, the imaging of the lesion area has uneven edges and a small gradient amplitude. The pixels in the area are more consistent with the characteristics of the lesion area, and the organs in the lung cross-section are approximately symmetrical relative to the central axis of the human body, while the shape of the lesion area is irregular and has poor symmetry. Based on the above characteristics, suspected lesion areas are screened to obtain the suspected target lesion areas in each CBCT image.
[0062] Preferably, in some implementations of the embodiments of the present invention, according to the degree of conformity to the lesion characteristics and the position distribution of edge pixels in the suspected lesion area, the specific method for obtaining the possibility that the suspected lesion area belongs to the lesion area is:
[0063] For any CBCT image in the lung tissue contour area If there are two edge pixels on the same horizontal line in the suspected lesion area, and the two edge pixels are located on both sides of the central axis of the CBCT image, the two edge pixels are recorded as valid pixel points of the suspected lesion area, and record the combination of the two edge pixel points as a valid pixel point combination;
[0064] For For any valid pixel point in any valid pixel point combination of the suspected lesion area, the Euclidean distance between the valid pixel point and the central axis of the CBCT image is recorded as the first distance of the valid pixel point; the absolute value of the difference between the first distances between two valid pixel points in the valid pixel point combination is recorded as the lesion characteristic distance of the valid pixel point combination;
[0065] The first The ratio between the mean value of the gradient amplitude of all edge pixels in the suspected lesion area and the mean value of the gradient amplitude of all edge pixels in all suspected lesion areas in the lung tissue contour area in the CBCT image is recorded as the first ratio; The ratio between the mean of the degree of conformity to lesion characteristics of all edge pixels of the first suspected lesion area and the mean of the degree of conformity to lesion characteristics of all edge pixels of all suspected lesion areas in the lung tissue contour area in the CBCT image is recorded as the second ratio; The number of all valid pixels in the suspected lesion area is The ratio of the number of all edge pixels in the suspected lesion area is recorded as the third ratio; The cumulative sum of the lesion feature distances of the effective pixel point combinations of the suspected lesion area is recorded as the first cumulative sum; the normalized value of the product of the first ratio, the second ratio, the third ratio and the first cumulative sum is recorded as the second cumulative sum. The possibility that the suspected lesion area belongs to the lesion area.
[0066] The specific formula is:
[0067]
[0068] In the formula, Indicates The possibility that the suspected lesion area belongs to the lesion area; Indicates The average value of the degree of conformity to lesion characteristics of all edge pixels in the suspected lesion area; It represents the mean value of the degree of conformity to lesion characteristics of all edge pixels of all suspected lesion areas in the lung tissue contour area in the CBCT image; Indicates The mean value of the gradient amplitude of all edge pixels in the suspected lesion area; It represents the mean value of the gradient amplitude of all edge pixels of all suspected lesion areas in the lung tissue contour area in the CBCT image; Indicates The number of all valid pixels in the suspected lesion area; Indicates The number of all edge pixels in the suspected lesion area; Indicates The total number of valid pixel combinations in the suspected lesion area; Indicates The first The lesion feature distance of the combination of effective pixels; represents the linear normalization function.
[0069] Preferably, in some implementations of the embodiments of the present invention, all suspected lesion regions are screened based on the possibility of belonging to the lesion region, and the specific method for obtaining the suspected target lesion region is:
[0070] Jordi The probability that the suspected lesion area belongs to the lesion area is greater than or equal to the lesion threshold parameter , will The suspected lesion area is recorded as the suspected target lesion area.
[0071] So far, all suspected target lesion areas in each CBCT image are obtained through the above method.
[0072] Step S003: according to the suspected target lesion areas in different CBCT images, all suspected target lesion pixel points are obtained; according to the difference between the degree of compliance with lesion characteristics and the possibility of belonging to the lesion area of the suspected target lesion pixel point in different CBCT images, the degree of compliance with artifact characteristics of each suspected target lesion pixel point is obtained; all lesion areas in the CBCT image are obtained based on the degree of compliance with artifact characteristics; according to the difference in the degree of compliance with artifact characteristics between each edge pixel point and the adjacent edge pixel points in the lesion area, all key lesion pixel points in the CBCT image are obtained.
[0073] 1. Obtain all lesion areas in CBCT images.
[0074] It should be noted that due to the physiological activities of the patient such as movement or breathing during the CBCT image acquisition process, artifacts may exist in the imaging results, so when the lesion area is determined by a single CBCT image, there may be interference from the artifact area. Artifacts are accidental phenomena in image acquisition and generally only exist in a single CBCT image. Therefore, the degree to which the pixel points at the same position in the artifact area in a single CBCT image conform to the characteristics of the lesion area is quite different from that in the other CBCT images. Due to the physiological activities of the human body, the position of the organs in the body is not completely consistent each time a CBCT image is acquired. Therefore, the suspected target lesion areas determined by multiple CBCT images are compared to obtain all lesion areas in the CBCT image.
[0075] Preferably, in some implementations of the embodiments of the present invention, the specific method for obtaining all suspected target lesion pixel points according to the suspected target lesion areas in different CBCT images is:
[0076] For any CBCT image, the grayscale values of all pixels in all suspected target lesion areas in the CBCT image are set to 1, and the other pixels are set to 0, to obtain a binary image of the CBCT image; the binary images of all CBCT images are subjected to a difference operation to obtain a CBCT difference image; for any pixel in the CBCT difference image, if the grayscale value of the pixel is not 0, the pixel is recorded as a suspected target lesion pixel;
[0077] Preferably, in some implementations of the embodiments of the present invention, according to the difference between the degree of compliance with lesion features and the possibility of belonging to the lesion area of the suspected target lesion pixel in different CBCT images, the specific method for obtaining the degree of compliance with artifact features of each suspected target lesion pixel is:
[0078] For any suspected target lesion pixel point, the ratio between the mean of the degree of conformity to lesion characteristics of all suspected target lesion pixels in all CBCT images and the mean of the degree of conformity to lesion characteristics of the suspected target lesion pixels in all CBCT images is recorded as the suspected lesion ratio; the ratio between the mean of the probability of belonging to the lesion area of the suspected lesion area to which all suspected target lesion pixels belong in all CBCT images and the mean of the probability of belonging to the lesion area of the suspected lesion area to which the suspected target lesion pixels belong in all CBCT images is recorded as the lesion ratio;
[0079] The difference between the degree of conformity of the suspected target lesion pixel point with the lesion feature in different CBCT images is used to obtain the degree of change of the lesion feature of the suspected target lesion pixel point;
[0080] The normalized value of the product of the lesion feature change degree of the suspected target lesion pixel, the lesion ratio and the suspected lesion ratio is used as the conforming artifact feature degree of the suspected target lesion pixel;
[0081] The specific formula is:
[0082]
[0083] In the formula, Indicates The degree to which the pixel of the suspected target lesion conforms to the artifact characteristics; It represents the mean value of the degree to which all suspected target lesion pixels in all CBCT images conform to the lesion characteristics; Indicates The average value of the degree to which the suspected target lesion pixel points in all CBCT images conform to the lesion characteristics; It represents the mean value of the probability that all suspected target lesion pixels belong to the suspected lesion area in all CBCT images; Indicates The mean value of the probability that the suspected target lesion pixel belongs to the suspected lesion area in all CBCT images; represents the total number of all CBCT images; Indicates The suspected target lesion pixel is at The degree of conformity with the lesion characteristics in each CBCT image; Indicates The suspected target lesion pixel is at The degree of conformity with the lesion characteristics in each CBCT image; Indicates taking the absolute value; represents the linear normalization function.
[0084] It should be noted that Indicates The degree of change in the lesion characteristics of the suspected target lesion pixel.
[0085] Preferably, in some implementations of the embodiments of the present invention, the specific method for acquiring all lesion areas in the CBCT image based on the degree of conformity with the artifact characteristics is:
[0086] Jordi The degree of conformity of the artifact characteristics of the suspected target lesion pixels is greater than or equal to the lesion threshold parameter , will The suspected target lesion pixel points are recorded as lesion pixels; the multiple regions formed by all lesion pixels in the CBCT difference image correspond to the multiple regions in the CBCT image, which are all recorded as lesion regions.
[0087] At this point, all lesion areas in the CBCT image are obtained.
[0088] 2. Obtain all key lesion pixels in the CBCT image.
[0089] It should be noted that in order to accurately match the lesion area determined in the CBCT image with the lesion area in the positioning CT, the key lesion pixels in the determined lesion area are screened to match the pixels in the CBCT image and the CT image to achieve regional registration; since the lesion area is irregular in shape and diverges from the middle to the outside, there is a situation where the edge curve has a large turning degree at a certain point on the regional contour; and due to the physiological activities of the human body, the determined lesion area has non-fixed pixels caused by pixels that are not affected by organ movement and position changes, and the degree to which the two pixels meet the artifact characteristics is quite different.
[0090] Preferably, in some implementations of the embodiments of the present invention, according to the difference in the degree of conformity to the artifact feature between each edge pixel point and the adjacent edge pixel points in the lesion area, the specific method for obtaining all key lesion pixel points in the CBCT image is:
[0091] In the lesion area The average of the degree of conformity of all pixels in the 8-neighborhood of the edge pixel is The absolute value of the difference between the degree of conformity to the artifact feature of the edge pixels is recorded as the first absolute value of the difference; The tangent vector of the edge pixel is The cosine value of the angle between the tangent vectors of edge pixels is recorded as The turning degree of the edge pixel point; the absolute value of the first difference and the The normalized value of the product of the turning degrees of the edge pixels is recorded as The probability of critical lesions at edge pixels;
[0092] The specific formula is:
[0093]
[0094] In the formula, Indicates the first The probability of critical lesions at edge pixels; Indicates The degree to which edge pixels conform to artifact characteristics; Indicates The average value of the degree of conformity to artifact characteristics of all pixels in the 8-neighborhood of an edge pixel; Indicates The tangent vector of the edge pixel is The cosine value of the angle between the tangent vectors of edge pixels; Indicates taking the absolute value; represents the linear normalization function.
[0095] Preferably, a key threshold parameter is preset , wherein this embodiment is based on This example is described as an example, and this embodiment is not specifically limited. Depends on the specific implementation situation;
[0096] Jordi The probability of critical lesions of edge pixels is greater than or equal to the critical threshold parameter , will The edge pixels are recorded as key lesion pixels.
[0097] At this point, all key lesion pixels in the CBCT image are obtained through the above method.
[0098] Step S004: According to the difference in the degree of conformity of the key lesion pixel points with all the pixels in the positioning CT image, the matching pixel points of each key lesion pixel point in the positioning CT image are obtained; based on the matching pixel points of the key lesion pixel points in the positioning CT image, the CBCT image and the positioning CT image are registered.
[0099] It should be noted that the key lesion pixel points obtained in the CBCT image are matched with the pixel points in the positioning CT image that have a high degree of similarity with the lesion characteristics, that is, the pixel points that are close in position in the image and have a similar degree of similarity with the lesion characteristics are matched.
[0100] Preferably, in some implementations of the embodiments of the present invention, according to the difference in the degree of conformity with the lesion characteristics between the key lesion pixel point and all the pixels in the positioning CT image, the specific method for obtaining the matching pixel point of each key lesion pixel point in the positioning CT image is:
[0101] For any pixel point in the localized CT image, The absolute value of the difference between the degree of conformity to the lesion characteristics of the key lesion pixel point and the degree of conformity to the lesion characteristics of the pixel point is recorded as the target difference; The Euclidean distance between the key lesion pixel and the pixel in the image is recorded as the target distance; the reciprocal of the product of the target difference and the target distance is recorded as The matching coefficient between a key lesion pixel point and the pixel point;
[0102] In the Among the matching coefficients between the key lesion pixel points and all the pixels in the positioning CT image, the pixel with the largest matching coefficient is recorded as According to the matching pixel points of each key lesion pixel in the positioning CT image, the CBCT image and the positioning CT image are registered.
[0103] See also Figure 2 , which shows a feature relationship flow chart for rapid CT image registration.
[0104] Through the above steps, a fast CT image registration method is completed.
[0105] The present invention also proposes a CT image rapid registration system, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the steps of a CT image rapid registration method in steps S001 to S004 are implemented.
[0106] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.
Claims
1. A method for rapid CT image registration, characterized in that: The method comprises the following steps: Acquire a positioning CT image and several CBCT images of the patient; According to the grayscale difference between each pixel point and the surrounding pixels in the lung tissue contour area in the CBCT image, the degree of conformity with the lesion characteristics of each pixel point in the lung tissue contour area is obtained; based on the degree of conformity with the lesion characteristics, all the pixels in the lung tissue contour area are screened to obtain the suspected lesion area; according to the degree of conformity with the lesion characteristics and the position distribution of the edge pixels in the suspected lesion area, the possibility of the suspected lesion area belonging to the lesion area is obtained; based on the possibility of belonging to the lesion area, all suspected lesion areas are screened to obtain the suspected target lesion area; According to the suspected target lesion areas in different CBCT images, all suspected target lesion pixel points are obtained; according to the difference between the degree of compliance with lesion characteristics and the possibility of belonging to the lesion area of the suspected target lesion pixel points in different CBCT images, the degree of compliance with artifact characteristics of each suspected target lesion pixel point is obtained; all lesion areas in the CBCT image are obtained based on the degree of compliance with artifact characteristics; according to the difference in the degree of compliance with artifact characteristics between each edge pixel point and the adjacent edge pixel points in the lesion area, all key lesion pixel points in the CBCT image are obtained; According to the difference in the degree of conformity of the key lesion pixel points with all the pixel points in the positioning CT image, the matching pixel points of each key lesion pixel point in the positioning CT image are obtained; based on the matching pixel points of the key lesion pixel points in the positioning CT image, the CBCT image and the positioning CT image are registered; The specific method of obtaining the degree of conformity of each pixel point in the lung tissue contour area with the lesion characteristics according to the grayscale difference between each pixel point and the surrounding pixels in the lung tissue contour area in the CBCT image includes: For any pixel point in the lung tissue contour area in any CBCT image, the grayscale difference between the pixel point and the pixels in its 8 neighborhoods is recorded as the grayscale change degree in the neighborhood of the pixel point; the product of the mean of the grayscale values of all pixels in the lung tissue contour area in the CBCT image and the grayscale change degree in the neighborhood of the pixel point is recorded as the first product; the normalized value of the ratio between the first product and the mean of the grayscale values of all pixels in the 8 neighborhoods of the pixel point is taken as the degree of compliance with the lesion characteristics of the pixel point; The specific method of obtaining the degree of conformity to artifact features of each suspected target lesion pixel point according to the difference between the degree of conformity to lesion features and the possibility of belonging to the lesion area of the suspected target lesion pixel point in different CBCT images includes: For any suspected target lesion pixel point, the ratio between the mean of the degree of conformity to lesion characteristics of all suspected target lesion pixels in all CBCT images and the mean of the degree of conformity to lesion characteristics of the suspected target lesion pixels in all CBCT images is recorded as the suspected lesion ratio; the ratio between the mean of the probability of belonging to the lesion area of the suspected lesion area to which all suspected target lesion pixels belong in all CBCT images and the mean of the probability of belonging to the lesion area of the suspected lesion area to which the suspected target lesion pixels belong in all CBCT images is recorded as the lesion ratio; The difference between the degree of conformity of the suspected target lesion pixel point with the lesion feature in different CBCT images is used to obtain the degree of change of the lesion feature of the suspected target lesion pixel point; The normalized value of the product of the lesion feature change degree of the suspected target lesion pixel, the lesion ratio and the suspected lesion ratio is used as the conforming artifact feature degree of the suspected target lesion pixel.
2. A CT image rapid registration method according to claim 1, characterized in that: The method of screening all pixel points in the lung tissue contour area based on the degree of conformity with the lesion characteristics to obtain the suspected lesion area includes the following specific methods: Preset a lesion threshold parameter For any pixel point in the lung tissue contour area in any CBCT image, if the degree of conformity of the pixel point to the lesion feature is greater than or equal to the lesion threshold parameter , the pixel point is recorded as a suspected lesion pixel point; and multiple regions formed by all the suspected lesion pixel points in the CBCT image are recorded as suspected lesion regions.
3. A method for rapid CT image registration according to claim 1, characterized in that: The method of obtaining the possibility that the suspected lesion area belongs to the lesion area according to the degree of compliance with the lesion characteristics and the position distribution of the edge pixels in the suspected lesion area includes the following specific methods: For any CBCT image in the lung tissue contour area If there are two edge pixels on the same horizontal line in the suspected lesion area, and the two edge pixels are located on both sides of the central axis of the CBCT image, the two edge pixels are recorded as valid pixel points of the suspected lesion area, and record the combination of the two edge pixel points as a valid pixel point combination; For For any valid pixel point in any valid pixel point combination of the suspected lesion area, the Euclidean distance between the valid pixel point and the central axis of the CBCT image is recorded as the first distance of the valid pixel point; the absolute value of the difference between the first distances between two valid pixel points in the valid pixel point combination is recorded as the lesion characteristic distance of the valid pixel point combination; The first The ratio between the mean value of the gradient amplitude of all edge pixels in the suspected lesion area and the mean value of the gradient amplitude of all edge pixels in all suspected lesion areas in the lung tissue contour area in the CBCT image is recorded as the first ratio; The ratio between the mean of the degree of conformity to lesion characteristics of all edge pixels of the first suspected lesion area and the mean of the degree of conformity to lesion characteristics of all edge pixels of all suspected lesion areas in the lung tissue contour area in the CBCT image is recorded as the second ratio; The number of all valid pixels in the suspected lesion area is The ratio of the number of all edge pixels in the suspected lesion area is recorded as the third ratio; The cumulative sum of the lesion feature distances of the effective pixel point combinations of the suspected lesion area is recorded as the first cumulative sum; the normalized value of the product of the first ratio, the second ratio, the third ratio and the first cumulative sum is recorded as the second cumulative sum. The possibility that the suspected lesion area belongs to the lesion area.
4. A method for rapid CT image registration according to claim 1, characterized in that: The specific method of obtaining all suspected target lesion pixels according to the suspected target lesion areas in different CBCT images includes: For any CBCT image, the grayscale values of all pixels in all suspected target lesion areas in the CBCT image are set to 1, and the other pixels are set to 0 to obtain a binary image of the CBCT image; the binary images of all CBCT images are subjected to a difference operation to obtain a CBCT difference image; for any pixel in the CBCT difference image, if the grayscale value of the pixel is not 0, the pixel is recorded as a suspected target lesion pixel.
5. A method for rapid CT image registration according to claim 4, characterized in that: The specific method of obtaining all lesion areas in the CBCT image based on the degree of compliance with the artifact characteristics includes: Jordi If the degree of conformity of the artifact characteristics of the suspected target lesion pixel point is greater than or equal to the preset threshold parameter, the first The suspected target lesion pixel points are recorded as lesion pixel points; Multiple regions formed by all lesion pixels in the CBCT difference image correspond to multiple regions in the CBCT image and are all recorded as lesion regions.
6. A method for rapid CT image registration according to claim 1, characterized in that: The method of obtaining all key lesion pixels in the CBCT image according to the difference in the degree of conformity to the artifact characteristics between each edge pixel in the lesion area and the adjacent edge pixel points includes the following specific methods: In the lesion area The average of the degree of conformity of all pixels in the 8-neighborhood of the edge pixel is The absolute value of the difference between the degree of conformity to the artifact feature of the edge pixels is recorded as the first absolute value of the difference; The tangent vector of the edge pixel is The cosine value of the angle between the tangent vectors of edge pixels is recorded as The turning degree of the edge pixel point; the absolute value of the first difference and the The normalized value of the product of the turning degrees of the edge pixels is recorded as The probability of critical lesions at edge pixels; Preset a key threshold parameter , The probability of critical lesions of edge pixels is greater than or equal to the critical threshold parameter , will The edge pixels are recorded as key lesion pixels.
7. A method for rapid CT image registration according to claim 1, characterized in that: The method of obtaining the matching pixel point of each key lesion pixel point in the positioning CT image according to the difference in the degree of conformity to the lesion characteristics between the key lesion pixel point and all the pixel points in the positioning CT image includes the following specific methods: For any pixel point in the localized CT image, The absolute value of the difference between the degree of conformity to the lesion characteristics of the key lesion pixel point and the degree of conformity to the lesion characteristics of the pixel point is recorded as the target difference; The Euclidean distance between the key lesion pixel and the pixel in the image is recorded as the target distance; the reciprocal of the product of the target difference and the target distance is recorded as The matching coefficient between a key lesion pixel point and the pixel point; In the Among the matching coefficients between the key lesion pixel points and all the pixels in the positioning CT image, the pixel with the largest matching coefficient is recorded as The matching pixels of the key lesion pixels.
8. A CT image rapid registration system, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of a method for rapid CT image registration as described in any one of claims 1 to 7 are implemented.
