English

CoDo: Contrastive Learning with Downstream Background Invariance for Detection

Computer Vision and Pattern Recognition 2022-05-11 v1 Artificial Intelligence

Abstract

The prior self-supervised learning researches mainly select image-level instance discrimination as pretext task. It achieves a fantastic classification performance that is comparable to supervised learning methods. However, with degraded transfer performance on downstream tasks such as object detection. To bridge the performance gap, we propose a novel object-level self-supervised learning method, called Contrastive learning with Downstream background invariance (CoDo). The pretext task is converted to focus on instance location modeling for various backgrounds, especially for downstream datasets. The ability of background invariance is considered vital for object detection. Firstly, a data augmentation strategy is proposed to paste the instances onto background images, and then jitter the bounding box to involve background information. Secondly, we implement architecture alignment between our pretraining network and the mainstream detection pipelines. Thirdly, hierarchical and multi views contrastive learning is designed to improve performance of visual representation learning. Experiments on MSCOCO demonstrate that the proposed CoDo with common backbones, ResNet50-FPN, yields strong transfer learning results for object detection.

Keywords

Cite

@article{arxiv.2205.04617,
  title  = {CoDo: Contrastive Learning with Downstream Background Invariance for Detection},
  author = {Bing Zhao and Jun Li and Hong Zhu},
  journal= {arXiv preprint arXiv:2205.04617},
  year   = {2022}
}

Comments

CVPR2022 workshop

R2 v1 2026-06-24T11:12:18.942Z