Dense Contrastive Learning for Self-Supervised Visual Pre-Training

Citations Over TimeTop 1% of 2021 papers

Abstract

To date, most existing self-supervised learning methods are designed and optimized for image classification. These pre-trained models can be sub-optimal for dense prediction tasks due to the discrepancy between image-level prediction and pixel-level prediction. To fill this gap, we aim to design an effective, dense self-supervised learning method that directly works at the level of pixels (or local features) by taking into account the correspondence between local features. We present dense contrastive learning (DenseCL), which implements self-supervised learning by optimizing a pairwise contrastive (dis)similarity loss at the pixel level between two views of input images.Compared to the baseline method MoCo-v2, our method introduces negligible computation overhead (only <1% slower), but demonstrates consistently superior performance when transferring to downstream dense prediction tasks including object detection, semantic segmentation and instance segmentation; and outperforms the state-of-the-art methods by a large margin. Specifically, over the strong MoCo-v2 baseline, our method achieves significant improvements of 2.0% AP on PASCAL VOC object detection, 1.1% AP on COCO object detection, 0.9% AP on COCO instance segmentation, 3.0% mIoU on PASCAL VOC semantic segmentation and 1.8% mIoU on Cityscapes semantic segmentation.Code and models are available at: https://git.io/DenseCL

Related Papers

• → Detecting Avocados to Zucchinis: What Have We Done, and Where Are We Going?(2013)88 cited
• → Object detection using improved YOLOv3-tiny based on pyramid pooling(2021)1 cited
• → Improving object detection via improving accuracy of object localization(2016)
• → Improved Regional Proposal Generation and Proposal Selection Method for Weakly Supervision Object detection(2023)