Research on Manipulation of Soft Tissue Based on 3D Vision*

摘要

The manipulation safety of robotic surgery is essential for tissue manipulation. The target object is often soft tissue during surgery, so the lack of estimation of deformation of them will bring substantial security risks. Therefore, we have studied how to manipulate soft tissue safely. Due to cost, size, accuracy, and sterilization requirements, obtaining haptic feedback directly through force sensors is still challenging. However, it is easier to get visual information through laparoscopic video. Therefore, we proposed a framework based on the 3D Vision to manipulate the soft tissue safely and conducted experiments to evaluate its feasibility. There are mainly three parts in this framework. First, obtain the 3D information of the scene and preprocess them. Second, use rigid and non-rigid registration to analyze the deformation of the soft tissue of interest between the current and previous frames. Third, define the security index(SI) and evaluate the validity and accuracy of SI. The experimental results show that the framework basically meets functional requirements.