Efficient and High-quality Ellipse Detection via Implicitly Excluding Most Useless Arc Groups and Enhancing Arc detection

Wencheng Wang, Zeyu Shen, Liebang Du, Mingyang Zhao

Keywords: ellipse detection, grid, edge extraction, edge-link

TL;DR: Improving the efficiency and accuracy of ellipse detection in images with implicit arc group excluding by grid based arc combining.

Abstract: Detecting ellipses from images is an fundamental problem in computer vision and pattern recognition, and plays an important role in many applications. This paper presents a new edge-link method for efficient and high-quality ellipse detection, where the two steps of edge-link methods are improved by our two presented novel measures respectively. The first is to adaptively adjust the search direction in linking edge pixels to generate arcs as consistently as possible. The second is to develop a novel measure for grouping arcs to check whether these arcs are from a same ellipse, which is by employing a grid to manage the arcs and designing a traversal path to visit grid cells continuously, through which most useless arc groups can be implicitly excluded for efficiency. This is different from existing methods that need explicitly check all possible arc groups. Based on these measures, we design an algorithm to detect ellipses as many as possible. Experimental results show that we can significantly improve both the accuracy and efficiency of ellipse detection, much superior to existing methods. Thus, we can significantly improve many applications.

Primary Area: Applications (e.g., vision, language, speech and audio, Creative AI)

Submission Number: 10034