Left/right hand segmentation in egocentric videos
详细信息 查看全文
- 作者：Alejandro Betancourt<sup> ; sup><sup>asup><sup> ; sup><sup>bsup><sup> ; sup><sup>a.betancourt@tue.nlsup> ; Pietro Morerio<sup>asup><sup> ; sup><sup>pmorerio@ginevra.dibe.unige.itsup> ; Emilia Barakova<sup>bsup><sup> ; sup><sup>e.i.barakova@tue.nlsup> ; Lucio Marcenaro<sup>asup><sup> ; sup><sup>lucio.marcenaro@unige.itsup> ; Matthias Rauterberg<sup>bsup><sup> ; sup><sup>g.w.m.Rauterberg@tue.nlsup> ; Carlo Regazzoni<sup>asup><sup> ; sup><sup>carlo@dibe.unige.itsup>
- 关键词：Hand-segmentation ; Hand-identification ; Egocentric vision ; First person vision
- 刊名：Computer Vision and Image Understanding
- 出版年：2017
- 出版时间：January 2017
- 年：2017
- 卷：154
- 期：Complete
- 页码：73-81
- 全文大小：2768 K
- 卷排序：154
文摘
Wearable cameras allow people to record their daily activities from a user-centered (First Person Vision) perspective. Due to their favorable location, wearable cameras frequently capture the hands of the user, and may thus represent a promising user-machine interaction tool for different applications. Existent First Person Vision methods handle hand segmentation as a background-foreground problem, ignoring two important facts: i) hands are not a single “skin-like” moving element, but a pair of interacting cooperative entities, ii) close hand interactions may lead to hand-to-hand occlusions and, as a consequence, create a single hand-like segment. These facts complicate a proper understanding of hand movements and interactions. Our approach extends traditional background-foreground strategies, by including a hand-identification step (left-right) based on a Maxwell distribution of angle and position. Hand-to-hand occlusions are addressed by exploiting temporal superpixels. The experimental results show that, in addition to a reliable left/right hand-segmentation, our approach considerably improves the traditional background-foreground hand-segmentation.