PDA: Proxy-based domain adaptation for few-shot image recognition

作者：

Highlights：

• Performing domain adaptation over few annotated samples for improved few-shot image recognition.

• Achieving task and domain transfer jointly.

• Proxy-based domain adaptation scheme without accessing source data.

• PDA optimizes the pre-trained representation and a novel few-shot classifier simultaneously.

• Achieving state-of-the-art results on multiple few-shot image recognition benchmarks.

摘要

•Performing domain adaptation over few annotated samples for improved few-shot image recognition.•Achieving task and domain transfer jointly.•Proxy-based domain adaptation scheme without accessing source data.•PDA optimizes the pre-trained representation and a novel few-shot classifier simultaneously.•Achieving state-of-the-art results on multiple few-shot image recognition benchmarks.

论文关键词：Few-shot image recognition,Domain adaptation,Few-shot learning,Transfer learning

论文评审过程：Received 2 November 2020, Revised 10 February 2021, Accepted 5 March 2021, Available online 23 March 2021, Version of Record 31 March 2021.

论文官网地址：https://doi.org/10.1016/j.imavis.2021.104164