Chenshu Chen — accepted AI-conference papers

2023

MaskBooster: End-to-End Self-Training for Sparsely Supervised Instance Segmentation

Shida Zheng, Chenshu Chen, Xi Yang, Wenming Tan

AAAI 2023technical

The present paper introduces sparsely supervised instance segmentation, with the datasets being fully annotated bounding boxes and sparsely annotated masks. A direct solution to this task is self-training, which is not fully explored for instance segmentation yet. In this paper, we propose MaskBoost…

Cited by 0 Source ↗PDF ↗Scholar ↗

2022

Dual Decoupling Training for Semi-supervised Object Detection with Noise-Bypass Head

Shida Zheng, Chenshu Chen, Xiaowei Cai, Tingqun Ye, Wenming Tan

AAAI 2022technical

Pseudo bounding boxes from the self-training paradigm are inevitably noisy for semi-supervised object detection. To cope with that, a dual decoupling training framework is proposed in the present study, i.e. clean and noisy data decoupling, and classification and localization task decoupling. In the…

Cited by 12 Source ↗PDF ↗Scholar ↗