Evaluation of 3D interest point detection techniques via human-generated ground truth
VISUAL COMPUTER, vol.28, pp.901-917, 2012 (SCI-Expanded, Scopus)
- Publication Type: Article / Article
- Volume: 28
- Publication Date: 2012
- Doi Number: 10.1007/s00371-012-0746-4
- Journal Name: VISUAL COMPUTER
- Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus
- Page Numbers: pp.901-917
- Keywords: 3D interest points, 3D salient points, 3D shape analysis
- Eskisehir Osmangazi University Affiliated: Yes
Abstract
In this paper, we present an evaluation strategy based on human-generated ground truth to measure the performance of 3D interest point detection techniques. We provide quantitative evaluation measures that relate automatically detected interest points to human-marked points, which were collected through a web-based application. We give visual demonstrations and a discussion on the results of the subjective experiments. We use a voting-based method to construct ground truth for 3D models and propose three evaluation measures, namely False Positive and False Negative Errors, and Weighted Miss Error to compare interest point detection algorithms.