Dual many-to-one-encoder-based transfer learning for cross-dataset human action recognition

作者：

Highlights：

• Proposed a new transfer-learning method for cross-dataset action recognition.

• A new dual many-to-one encoder method for feature extraction across action datasets.

• Achieved over 10% increase in recognition accuracy over recent work.

摘要

•Proposed a new transfer-learning method for cross-dataset action recognition.•A new dual many-to-one encoder method for feature extraction across action datasets.•Achieved over 10% increase in recognition accuracy over recent work.

论文关键词：Cross-dataset,Action recognition,Neural network,Transfer learning,Domain adaptation

论文评审过程：Received 16 September 2015, Revised 10 December 2015, Accepted 2 January 2016, Available online 14 January 2016, Version of Record 10 November 2016.

论文官网地址：https://doi.org/10.1016/j.imavis.2016.01.001