Accurate detection of ellipses with false detection control at video rates using a gradient analysis. (September 2018)

• Record Type: Journal Article
• Authors: Dong, Huixu; Prasad, Dilip K.; Chen, I-Ming
• Is Part Of: Pattern recognition. Volume 81(2018:Sep.)
• Page Start: 112
• Page End: 130
• Publication Date: 2018-09
• Subjects: Ellipse detection -- Geometric approach -- Gradient analysis -- Centre estimation -- Arc classification; Pattern perception -- Periodicals; Perception des structures -- Périodiques; Patroonherkenning; 006.4
• DOI: 10.1016/j.patcog.2018.03.023
• Languages: English
• ISSNs: 0031-3203

Abstract

Highlights: This paper presents an ellipse detection method that combines the advantages of arc extraction and arc grouping to guarantee the effectiveness of ellipse detection and optimizes the computation cost. In the step of smooth arc extraction, we propose a novel approach of identifying the precise splitting points (sudden changes) in order to achieve better segmentations from curves to smooth arcs that may belong to ellipses. A coarse search for sudden changes is first performed with a big range, and then such points are determined with a finer scope. We present a novel method to estimate the ellipse centre by an iterative mean-shift clustering algorithm, which improves its robustness to noise and obtains a more precise centre comparing the existing methods that determine ellipse centres. We adopt the ratio of half of the circumference of the bounding box enclosing an arc and the sum of the semi-axes lengths to measure the integrity of ellipse to improve the detection accuracy. We propose a new approach of false determination control to determine detection results based on the intrinsic geometric attribute of ellipse expressed by a mathematical model, which avoids false detections effectively. Abstract: Accurate ellipse detection in image streams at real-time execution is an open challenge. We present a novel fast and robust ellipse detection method. The method adopts arcs selection, smart grouping, and repeated utilization of gradient information to significantly reduce the computations otherwise needed without compromising the detection effectiveness. Geometric properties calculable with few computations, such as arc smoothness, relative placement of curves, and region of confidence for ellipse centres, are utilized for this purpose. An exhaustive sensitivity analysis of the method's control parameters has been performed. It reveals range of values that support consistent performance over diverse challenging datasets with complex background, multiple differently sized ellipses, and occluded, overlapping ellipses. The method's performance is compared with six state-of-the-art detectors over four diverse datasets. Among all the tested methods, the proposed method demonstrates the best balance between detection effectiveness (the best or the second best F-measure scores) and computation time (>40 Hz) across all the datasets.