Rib recognition method and system for intraoperative imaging

Abstract

The present application provides a rib recognition method and system for intraoperative images. The rib recognition method includes: inputting the intraoperative images into the first segmentation model to obtain the first intraoperative bone segmentation images. The intraoperative images are included in the chest surgery process. A partial image of the chest taken in . Inputting the preoperative images into the second segmentation model to obtain preoperative bone segmentation images with classification results, wherein the preoperative images include images of the chest taken before thoracic surgery, the first intraoperative bone segmentation images and the preoperative bone segmentation images Part of the image corresponds in . mapping the first intraoperative bone segmentation image to the preoperative bone segmentation image to obtain a second intraoperative bone segmentation image, wherein the second intraoperative bone segmentation image includes the first intraoperative bone segmentation image and the preoperative bone segmentation image Classification results for ribs at the same position in . This application combines the intraoperative image with the preoperative image, effectively uses the preoperative image to plan the intraoperative plan, and assists the doctor in positioning the puncture point during the operation.

Description

technical field [0001] This application relates to the field of deep learning technology, in particular to a rib recognition method and system for intraoperative images. Background technique [0002] In the percutaneous puncture positioning operation, the rib is a very important organ. In order to effectively avoid the area where the rib is located during the puncture, it is necessary to segment and classify the rib. However, in the process of percutaneous puncture positioning, in order to ensure precise avoidance of ribs, multiple scans are usually performed around the target area to obtain images near the target area. However, multiple scans will not only bring multiple radiation effects to the patient, but also make it difficult to classify and identify ribs due to the lack of complete images of the lung area. Contents of the invention [0003] In view of this, the embodiment of the present application provides a method and system for identifying ribs in an intraoperat...

AI Technical Summary

Problems solved by technology

However, multiple scans will not only bring multiple radiation effects to patients, but also make it difficult to classify and identify ribs due to the lack of complete images of the lung area.

Images