Region-based dropout with attention prior for weakly supervised object localization

作者：

Highlights：

• We propose a new WSOL method, RDAP, that induces a model to learn the less discriminative parts of an object.

• The proposed method is more robust to the selection of hyperparameters, compared to previous WSOL methods.

• We achieved state-of-the-art performances on two popular datasets with four architectures.

摘要

•We propose a new WSOL method, RDAP, that induces a model to learn the less discriminative parts of an object.•The proposed method is more robust to the selection of hyperparameters, compared to previous WSOL methods.•We achieved state-of-the-art performances on two popular datasets with four architectures.

论文关键词：Deep learning,Object localization,Weakly supervised learning,Region-based dropout,Attention prior

论文评审过程：Received 11 May 2020, Revised 20 February 2021, Accepted 14 March 2021, Available online 19 March 2021, Version of Record 26 March 2021.

论文官网地址：https://doi.org/10.1016/j.patcog.2021.107949