Training object detectors from few weakly-labeled and many unlabeled images

作者：

Highlights：

• A novel method to train detector by few weakly-labeled images and lots of unlabeled images.

• The features extracted from the labeled images by a pretrained classifier are used to label unsupervised images.

• The proposed approaches result in competitive performance with state-of-the-art weakly supervised object detectors.

摘要

•A novel method to train detector by few weakly-labeled images and lots of unlabeled images.•The features extracted from the labeled images by a pretrained classifier are used to label unsupervised images.•The proposed approaches result in competitive performance with state-of-the-art weakly supervised object detectors.

论文关键词：Object detection,Weakly-supervised learning,Semi-supervised learning,Unlabelled set

论文评审过程：Received 28 October 2020, Revised 24 June 2021, Accepted 4 July 2021, Available online 5 July 2021, Version of Record 12 July 2021.

论文官网地址：https://doi.org/10.1016/j.patcog.2021.108164