Weakly-supervised Instance Segmentation via Class-agnostic Learning with Salient Images

作者：Xinggang Wang, Jiapei Feng, Bin Hu, Qi Ding, Longjin Ran, Xiaoxin Chen, Wenyu Liu · 年份：2021 · DOI：10.1109/cvpr46437.2021.01009 · 被引用次数：35 · 研究领域：Advanced Neural Network Applications、Visual Attention and Saliency Detection、Domain Adaptation and Few-Shot Learning

Humans have a strong class-agnostic object segmentation ability and can outline boundaries of unknown objects precisely, which motivates us to propose a box-supervised class-agnostic object segmentation (BoxCaseg) based solution for weakly-supervised instance segmentation. The BoxCaseg model is jointly trained using box-supervised images and salient images in a multi-task learning manner. The fine-annotated salient images provide class-agnostic and precise object localization guidance for box-supervised images. The object masks predicted by a pretrained BoxCaseg model are refined via a novel merged and dropped strategy as proxy ground truth to train a Mask R-CNN for weakly-supervised instance segmentation. Only using 7991 salient images, the weakly-supervised Mask R-CNN is on par with fully-supervised Mask R-CNN on PASCAL VOC and significantly outperforms previous state-of-the-art box-supervised instance segmentation methods on COCO. The source code, pretrained models and datasets are available at https://github.com/hustvl/BoxCaseg.