CN-121414737-B - Human body contour sketching algorithm system based on CT image

CN121414737BCN 121414737 BCN121414737 BCN 121414737BCN-121414737-B

Abstract

The invention discloses a human body contour sketching algorithm system based on CT images, which relates to the technical field of medicine, and comprises the steps of converting CT image data into an RAS human body coordinate system, segmenting the converted CT image data by a threshold segmentation method to obtain basic body mask data, removing body films and mechanical structure contour lines in the basic body mask data through morphological opening operation, reading scanning body position data from head information, judging to be in a prone position or a supine position, determining the position of a bed plate relative to a human body, filling a missing region of a body region according to the scanning body position by adopting a differentiation strategy, removing bed plate interference from the body region subjected to filling treatment, and filling a global cavity into the body region subjected to bed plate interference removal to obtain a complete human body contour sketching result. The method and the device remarkably improve the accuracy and the integrity of sketching and realize effective recognition and accurate removal of the stubborn bed board trace.

Inventors

GU YAN
HU WEI
ZHENG CHAO

Assignees

海创未来(杭州)医疗科技有限公司

Dates

Publication Date: 20260508
Application Date: 20251225

Claims (5)

1. A human body contour delineation algorithm system based on CT images, comprising: the preprocessing module is used for converting the CT image data into an RAS human body coordinate system, dividing the converted CT image data by a threshold segmentation method and obtaining basic body mask data; The body film removing module is used for removing body films and mechanical structure contour lines in the basic body mask data through morphological opening operation; the body position information extraction module reads scanned body position data from head information, judges the scanned body position data to be a prone position or a supine position, and determines the position of the bed board relative to a human body; The local region filling module is used for filling the missing region of the body region according to the scanning body position by adopting a differentiation strategy; the bed board removing module is used for removing bed board interference on the body area after the filling treatment is completed; The method comprises the steps of performing a bed board removal operation on supine position data, dividing the bed board into a non-refractory bed board and a refractory bed board, performing two-dimensional communication domain screening on the non-refractory bed board after a structure is separated from a body area through a corrosion operation with a core of 7, performing a three-dimensional maximum communication domain retaining operation, and finally performing an expansion operation with the core of 7 to restore the original mask size, scanning the refractory bed board from the bed board side to the body area in each layer of CT image, recording the mask pixel change ratio R of each row, sequencing the change ratio of each row to obtain a sequence R, removing the influence of reducing abnormal values by the front 5% and the rear 5%, and taking an average value to obtain the average change condition of mask pixels at the side of the CT image of the layer The obtained And the maximum in the sequence If the comparison is more than twice, the boundary between the refractory bed board and the body is considered to be the boundary for cutting the bed board, otherwise, the bed board is not existed and is not treated; ; ; Wherein, the For the previous line of mask pixels in the current layer CT image, For the CT image of the current layer the number of pixels of the current layer mask; And the global region filling module is used for filling global cavities in the body region from which the interference of the bed board is removed, so that a complete human body contour sketching result is obtained.
2. The human body contour sketching algorithm system based on the CT image according to claim 1, wherein a standardized function built in a SIMPLEITK library is adopted in the preprocessing module to perform coordinate system conversion on the CT image data.
3. The human body contour sketching algorithm system based on the CT image according to claim 1, wherein an optimal threshold value of threshold segmentation in morphological opening operation is determined by adopting an Ojin method in the body film removing module.
4. The human body contour sketching algorithm system based on the CT image according to claim 1, wherein the differentiated filling strategy in the local area filling module is specifically that prone position data is subjected to open operation with a kernel size of 5 to remove bed board interference and then body front side filling, and supine position data is subjected to body front side filling directly.
5. The CT image-based human body contour delineation algorithm system of claim 4, wherein the filling specifically comprises performing an expansion operation with a 9-size kernel, filling a blank portion of the closed region by a flooding filling algorithm, and restoring the original contour size by a 9-size kernel.

Description

Human body contour sketching algorithm system based on CT image Technical Field The invention relates to the technical field of medicine, in particular to a human body contour sketching algorithm system based on CT images. Background In the radiotherapy treatment process, the whole body region in the CT image is accurately segmented, so that the body boundary of a patient can be clearly defined. This step is critical for the accurate positioning of the radiotherapy plan, and ensures that the radiation is accurately irradiated to the predetermined area, effectively preventing the radiation from diffusing to other non-target parts of the body. In addition, the whole body segmentation technology is also helpful for identifying the posture, the body position and the body contour of the patient, and provides firm support for accurate positioning in the treatment process. Currently, body region segmentation algorithms are mainly divided into two main categories, traditional image segmentation algorithms and image segmentation algorithms based on deep learning. In conventional image segmentation algorithms, either thresholding, region segmentation, or edge-based segmentation methods, are highly dependent on the apparent difference between the foreground region (i.e., whole-body region) and the background region. However, in the actual CT image, the gray scale and texture of the bed board of the CT machine tool are often similar to those of the body tissue, and in addition, the problems of incomplete body contour and the like may exist in the positioning body film used in the radiotherapy process, as shown in fig. 1, all these factors affect the performance of the conventional algorithm to different degrees. Likewise, image segmentation algorithms based on deep learning also face challenges in dealing with different scan positions, patient posture differences, and couch board to body contact conditions. To cope with these complexities, it is often necessary to collect a large amount of training data with accurate labels. However, the collection of body whole tags is more complex and time consuming than tag data for organs or target areas in a radiotherapy scenario, which presents a significant challenge for data collection due to the broader scope and more detail involved. Therefore, how to accurately identify the posture, the body position and the body contour of the patient in the CT image is a problem to be solved by those skilled in the art. Disclosure of Invention In view of the above, the present invention provides a human body contour sketching algorithm system based on CT images to solve the problems in the background art. In order to achieve the above purpose, the present invention adopts the following technical scheme: A human body contour delineation algorithm system based on CT images, comprising: the preprocessing module is used for converting the CT image data into an RAS human body coordinate system, dividing the converted CT image data by a threshold segmentation method and obtaining basic body mask data; The body film removing module is used for removing body films and mechanical structure contour lines in the basic body mask data through morphological opening operation; the body position information extraction module reads scanned body position data from head information, judges the scanned body position data to be a prone position or a supine position, and determines the position of the bed board relative to a human body; The local region filling module is used for filling the missing region of the body region according to the scanning body position by adopting a differentiation strategy; the bed board removing module is used for removing bed board interference on the body area after the filling treatment is completed; And the global region filling module is used for filling global cavities in the body region from which the interference of the bed board is removed, so that a complete human body contour sketching result is obtained. Preferably, the preprocessing module adopts a standardized function built in SIMPLEITK libraries to perform coordinate system conversion on the CT image data. Preferably, the body membrane removing module adopts an oxford method to determine an optimal threshold value of threshold segmentation in morphological opening operation. Preferably, the differential filling strategy in the local area filling module is specifically that prone position data is subjected to open operation with a kernel size of 5 to remove interference of a bed board, and then body front side filling is performed, and supine position data is directly subjected to body front side filling. Preferably, the filling specifically comprises the steps of performing expansion operation on the front half area of the body by adopting a check with the size of 9, filling the blank part of the closed area by adopting a water-diffusion filling algorithm, and then performing corrosion operation on the core with th