Accurate Perception of Manipulation Force Makes Vascular Intervention Safer

Abstract

Manipulation force plays an important role in ensuring the safety of vascular intervention. Due to its extreme slightness, few studies have achieved accurate perception. In this study, a vision-based perception system is developed to obtain the manipulation force during guidewire delivery. In the system, an analysis framework is designed with the Hough transform and a fully connected neural network to establish the accurate transformation from visual image to haptic force. Experimental results indicate that the perception system achieves a minimal error (less than 0.1 N) in predicting the true force. It provides an effective solution for realtime monitoring dynamic manipulation in the context of vascular intervention. The proposed method also holds great potential for applications that require accurate force perception and control in various fields.