Comparative study of feature localization methods for image matching in endoscopic images

Abstract

The purpose of this work is to determine which is the best general method of feature localization for image matching in endoscopy images. In order to accomplish this, we conduct an exhaustive analysis of ten well-known feature detectors, descriptors, and deep learning algorithms, such as: SIFT, FAST, SURF, ORB, BRIEF, BRISK, FREAK, HARRIS, DFM, and LOFTR. The programming code for applying the aforementioned algorithms is developed in Python. The analysis is performed across six challenging medical datasets: cardiorespiratory endoscopy, human laparoscopy, bronchoscopy, gastroscopy, animal laparoscopy, and pig endoscopy. This framework is highly diverse, containing a variety of textures, camera motions, tissue deformations, and visual barriers. To determine which technique is the best on average, we perform a qualitative analysis of the inliers and a quantitative analysis using the following metrics: number of keypoints, number of matches, number of inliers, computational cost, and sparsity of the inliers among them. Two different scenarios are considered when analyzing algorithms: sequential and template-based scenarios. Furthermore, we examine how those features may be exploited in the reconstruction of 3D shapes from visual cues.