From handcrafted to learned representations for human action recognition: A survey
作者:
Highlights:
• We provide a comprehensive stud over the state-of-the-art action representations.
• Deep learning-based representations are compared to handcrafted representations.
• Pros and cons of current deep learning-based approaches are discussed.
摘要
•We provide a comprehensive stud over the state-of-the-art action representations.•Deep learning-based representations are compared to handcrafted representations.•Pros and cons of current deep learning-based approaches are discussed.
论文关键词:Human action recognition,Handcrafted features,Deep learning,Convolutional neural network,Dictionary learning
论文评审过程:Received 15 September 2015, Revised 7 May 2016, Accepted 16 June 2016, Available online 27 June 2016, Version of Record 10 November 2016.
论文官网地址:https://doi.org/10.1016/j.imavis.2016.06.007