Few-shot learning-based human activity recognition

作者：

Highlights：

• We propose a novel few-shot learning scheme for human activity recognition.

• We propose a general framework to measure cross-domain class-wise relevance.

• Negative transfer alleviated.

• We design a deep learning framework for human activity recognition.

摘要

•We propose a novel few-shot learning scheme for human activity recognition.•We propose a general framework to measure cross-domain class-wise relevance.•Negative transfer alleviated.•We design a deep learning framework for human activity recognition.

论文关键词：Human activity recognition,Few-shot learning,Knowledge transfer,Cross-domain class-wise relevance,Deep learning

论文评审过程：Received 31 March 2019, Revised 9 June 2019, Accepted 30 June 2019, Available online 4 July 2019, Version of Record 19 July 2019.

论文官网地址：https://doi.org/10.1016/j.eswa.2019.06.070