Weak Supervision with Arbitrary Single Frame for Micro- and Macro-expression Spotting

arXiv:2403.14240v1 Announce Type: new
Abstract: Frame-level micro- and macro-expression spotting methods require time-consuming frame-by-frame observation during annotation. Meanwhile, video-level spotting lacks sufficient information about the location and number of expressions during training, resulting in significantly inferior performance compared with fully-supervised spotting. To bridge this gap, we propose a point-level weakly-supervised expression spotting (PWES) framework, where each expression requires to be annotated with only one random frame (i.e., a point). To mitigate the issue of sparse label distribution, the prevailing solution …

abstract annotation arxiv bridge cs.cv gap information location macro micro observation performance supervision training type video