Selective Weakly Supervised Human Detection under Arbitrary Poses

详细信息    查看全文

• 作者：Yawei Cai^a ; ^bAuthor Vitae ; Xiaosong Tan^a ; ^bAuthor Vitae ; Xiaoyang Tan^a ; ^b ; ^{x.tan@nuaa.edu.cn}Author Vitae
• 关键词：Weakly supervised learning ; Human detection ; Selective Weakly Supervised Detection (SWSD) ; Multi-instance learning (MIL)
• 刊名：Pattern Recognition
• 出版年：2017
• 出版时间：May 2017
• 年：2017
• 卷：65
• 期：Complete
• 页码：223-237
• 全文大小：2768 K
• 卷排序：65

文摘

We propose a novel Selective Weakly Supervised Detection method which outperforms the previous state-of-the-art methods. We annotate a new large-scale data set called LSP/MPII-MPHB (Multiple Poses Human Body) for human body detection. We identify an easily ignored pitfall of the Noisy-OR model in MIL, which can significantly reduce the training efficiency of a MIL algorithm. We present a comprehensive and in-depth empirical study of the weakly supervised MIL method.