Guide-Wire Detecting Based on Speeded up Robust Features for Percutaneous Coronary Intervention

Abstract

Percutaneous coronary intervention (PCI) is a type of endovascular surgery. In the PCI procedure, guide-wire threading under the monitoring of X-ray videos is a vital step widely used to treat narrowing stenosis of a coronary artery. Detection of guide-wire in X-ray videos is not a trivial task because guide-wire has various shapes, and the signal to noise rate is pretty low. Besides, some anatomical skeleton contours are similar to guide-wires. Therefore, it urgently needs accuracy and robust method. In this research, we present a fast and robust guide-wire detection method we offer a fast and robust guide-wire detection method, speeded up robust features (SURF) is applied to locate the tip of guide-wire in various shapes and situations. Our approach was evaluated by testing on 18 X-ray sequence images, total 1073 frames (50 frames for training and 1023 frames for testing). The detection accuracy is 92.7% with 20 fps speed that shows a promising result for guide-wires detection.