Unsupervised Object-Level Representation Learning from Scene Images
arXiv:2106.11952
Abstract
Contrastive self-supervised learning has largely narrowed the gap to supervised pre-training on ImageNet. However, its success highly relies on the object-centric priors of ImageNet, i.e., different augmented views of the same image correspond to the same object. Such a heavily curated constraint becomes immediately infeasible when pre-trained on more complex scene images with many objects. To overcome this limitation, we introduce Object-level Representation Learning (ORL), a new self-supervised learning framework towards scene images. Our key insight is to leverage image-level self-supervised pre-training as the prior to discover object-level semantic correspondence, thus realizing object-level representation learning from scene images. Extensive experiments on COCO show that ORL significantly improves the performance of self-supervised learning on scene images, even surpassing supervised ImageNet pre-training on several downstream tasks. Furthermore, ORL improves the downstream performance when more unlabeled scene images are available, demonstrating its great potential of harnessing unlabeled data in the wild. We hope our approach can motivate future research on more general-purpose unsupervised representation learning from scene data.
NeurIPS 2021. Project page: https://www.mmlab-ntu.com/project/orl/ Code: https://github.com/Jiahao000/ORL
References in corpus (8)
- Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
- Exploring Simple Siamese Representation Learning
- Self-supervised Pretraining of Visual Features in the Wild
- Self-EMD: Self-Supervised Object Detection without ImageNet
- Dense Contrastive Learning for Self-Supervised Visual Pre-Training
- Mine Your Own vieW: Self-Supervised Learning Through Across-Sample Prediction
- Spatially Consistent Representation Learning
- Instance Localization for Self-supervised Detection Pretraining