Multimodal Failure Matching Point Based Motion Object Saliency Detection for Unconstrained Videos. Issue 1 (31st December 2022)

• Record Type: Journal Article
• Authors: Qian, Jiang; Wei, Jingkang; Chen, Hui; Chen, Gongping
• Is Part Of: Applied artificial intelligence. Volume 36:Issue 1(2022)
• Publication Date: 2022-12-31
• Subjects: Artificial intelligence -- Periodicals; 006.3
• DOI: 10.1080/08839514.2022.2110695
• Languages: English
• ISSNs: 0883-9514

Abstract

ABSTRACT: Inspired by classical feature descriptors in motion matching, this paper proposes a multimodal failure matching point collection method, which is defined as FMP. FMP is, in fact, a collection of unstable features with a low matching degree in the conventional matching task. Based on FMP, a novel model for the saliency detection of motion object is developed. Models are evaluated on the DAVIS and SegTrackv2 datasets and compared with recently advanced object detection algorithms. The comparison results demonstrate the availability and effectiveness of FMP in the detection of motion object saliency.