Trajectory planning for robot-assisted laminectomy decompression based on CT images

Abstract

Laminectomy decompression is one of the most complex spinal operations, with a high surgical risk and surgeon fatigue. The introduction of robots into surgery is expected to effectively solve these problems, but the complex and time-consuming grinding planning hinders the research and application of robot-assisted laminectomy. This paper proposes a robot grinding path automatic generation method for this operation to simplify the planning process. First, a neural network is designed to obtain the central positions of laminae in a CT image. Around the laminar center, a series of sparse robotic motion control points are obtained and adjusted based on bone surface. Simulation experiments based on some spine CT datasets indicate that the proposed method can effectively generate a reasonable planned path from spine CT images.