GHT-based associative memory learning and its application to Human action detection and classification
作者：
Highlights：
• GHT-based associative memory is proposed to recover missing parts of a video object from parts of detectable patches.
• The approach is robust even when we apply it to detect video action objects from a video clip with a cluttered background.
• We propose an automatic training procedure for generating efficient and discriminative associative memory models.
• Techniques to smartly select the patches and synthesize the target action shape are implemented.
摘要
•GHT-based associative memory is proposed to recover missing parts of a video object from parts of detectable patches.•The approach is robust even when we apply it to detect video action objects from a video clip with a cluttered background.•We propose an automatic training procedure for generating efficient and discriminative associative memory models.•Techniques to smartly select the patches and synthesize the target action shape are implemented.
论文关键词：Action object shapes,Generalized Hough transform,Associative memory,Hypergraph,Human action detection and recognition
论文评审过程：Received 5 November 2012, Revised 21 February 2013, Accepted 26 March 2013, Available online 24 April 2013.
论文官网地址：https://doi.org/10.1016/j.patcog.2013.03.027
