Pedicle screw operation path automatic planning method based on deep learning network

Abstract

The invention relates to the field of medicine, in particular to a spinal pedicle screw operation path planning method based on deep learning. The method comprises the following steps: expressing a spinal pedicle screw operation path as a linear operation path, and defining an operation entry point and an operation direction point; establishing an operation path planning data set comprising a spine segmentation data set and an operation path key point data set; designing a spine segmentation network in a mode of jointly supervising the network by adopting an encoder-decoder structure for fivetimes of down-sampling, Dice loss and softmax loss; combining a convolution network and a full connection network, and in a manner that L1 loss and mean square error loss jointly supervise the network, designing a surgical path point positioning network; and automatically segmenting the spinal CT image by adopting the trained network, automatically positioning the key points of the operation path,reconstructing the operation path through the key points, and evaluating the planning of the operation path by adopting two modes of subjective evaluation and objective evaluation. According to the method, the screw entering path of the spinal pedicle screw operation can be automatically planned.

Description

technical field [0001] The present invention relates to the medical field, in particular to a deep learning-based spinal surgery path planning method. Background technique [0002] Surgical path planning is a key core step in realizing intelligent orthopedic surgery (especially image-guided computer-assisted or robot-assisted orthopedic surgery). Traditional surgical path planning methods mostly rely on the experience of the surgeon (doctor). Factors such as the physiological state and subjective judgment of the surgeon significantly affect the quality of path planning, and the stability of surgical planning results and operational efficiency cannot be effectively guaranteed. Therefore, seeking a more efficient and intelligent surgical path planning method is becoming a research and application hotspot in intelligent orthopedic surgery. [0003] The early path planning methods for spinal surgery were mainly manual or semi-automatic interactive operations performed by the op...

AI Technical Summary

Problems solved by technology

The accuracy of the existing SSM technology is still difficult to meet the accuracy requirements of path planning for spinal robotic surgery

Images