Intelligent tracking method and system for skin tumor precise surgery based on machine vision

Abstract

The invention discloses an intelligent tracking method and system for a skin tumor precise surgery based on machine vision. The method comprises the following steps of conducting full-process high-definition shooting on the extraction process of a subcutaneous tissue slice and the re-excision process of a corresponding area of the subcutaneous tissue slice, and transmitting a shot video to a visual computing unit in real time; detecting and tracing a surgical blade by the visual computing unit according to the shot video, wherein specifically, the central point of a movement trajectory of thesurgical blade is calculated, then a corresponding surgical area is marked, a re-excision area is matched with a subcutaneous tissue slice extraction area, and a matching result is transmitted to a human-machine interaction unit; the human-machine interaction unit displays the matching result of the visual computing unit to a user, and simultaneously receives an operation instruction of the user and transmits the instruction to the visual computing unit. By means of the intelligent tracking method and system, the subcutaneous tissue slice extraction area and the re-excision area are marked bytracking the movement trajectory of the surgical blade, and the system is expected to replace a human assistant in large-area skin tumor excision surgeries.

Description

technical field [0001] The present invention relates to the field of intelligent assistance for precision surgery, in particular to a machine vision-based intelligent tracking method and system for skin tumor precision surgery. Background technique [0002] At present, in skin tumors, especially large-area skin tumor resection, specialized human assistants are required to identify, memorize, and match the subcutaneous tissue section extraction areas. The intelligent tracking technology and method of skin tumor precision surgery based on machine vision can complete the work of human assistants by tracking the movement trajectory of the surgical blade. In recent years, with the rise of deep learning, especially convolutional neural networks, a series of target detection and segmentation models based on convolutional neural networks have emerged, including R-CNN, FastR-CNN, FasterR-CNN, and MaskR-CNN. The current target tracking technology basically uses a border to identify t...

AI Technical Summary

Problems solved by technology

The current target tracking technology basically uses a border to identify the target object, and mainly adopts the two-stage strategy of "first frame target detection + subsequent frame target tracking", which is suitable for most application scenarios; however, in In some scenes that require high tracking accuracy, if the relative position of the target object in the tracking bounding box varies greatly, the tracking accuracy requirements will not be met.

Images