ㅇ 제목 : Deep Weakly Supervised Temporal Action Localization

ㅇ 일시 : 2018년 5월 24일(목) 10:00 - 12:00

ㅇ 장소 : ETRI 융합기술연구생산센터 224호실

ㅇ 강사 : 서울대 한보형 교수

ㅇ 요약 : 

The success of deep learning in computer vision is partly attributed to the construction of large-scale labeled datasets such as ImageNet. However, computer vision problems often require substantial human efforts and interventions to obtain accurate annotations due to dynamic aspects of class labels, needs for pixel-level labeling, and annotation ambiguities. Hence, collecting high quality large-scale labeled datasets is very time consuming and even infeasible. This talk addresses challenges and approaches related to data deficiency issue in the context of computer vision including object detection, semantic segmentation, and action localization, and discusses a solution about weakly supervised temporal action localization problem.