Classification and weakly supervised pain localization using multiple segment representation

作者：

Highlights：

• Target problem of pain classification and localization using weakly-labeled pain videos

• Algorithm combines multiple instance learning with multiple segment representation.

• Rigorous experiments show that our algorithm achieves state-of-the-art performance.

• Empirically evaluate the contributions of different components of our algorithm

• Visualize discriminative facial patches, as learned by our algorithm

摘要

•Target problem of pain classification and localization using weakly-labeled pain videos•Algorithm combines multiple instance learning with multiple segment representation.•Rigorous experiments show that our algorithm achieves state-of-the-art performance.•Empirically evaluate the contributions of different components of our algorithm•Visualize discriminative facial patches, as learned by our algorithm

论文关键词：Emotion classification,Action classification,Pain,Temporal segmentation,Bag of Words,Weakly supervised learning,Boosting,Bagging

论文评审过程：Received 2 June 2013, Revised 10 December 2013, Accepted 21 February 2014, Available online 4 March 2014.

论文官网地址：https://doi.org/10.1016/j.imavis.2014.02.008