Learning Video Object Segmentation from Unlabeled Videos
arXiv:2003.05020
Abstract
We propose a new method for video object segmentation (VOS) that addresses object pattern learning from unlabeled videos, unlike most existing methods which rely heavily on extensive annotated data. We introduce a unified unsupervised/weakly supervised learning framework, called MuG, that comprehensively captures intrinsic properties of VOS at multiple granularities. Our approach can help advance understanding of visual patterns in VOS and significantly reduce annotation burden. With a carefully-designed architecture and strong representation learning ability, our learned model can be applied to diverse VOS settings, including object-level zero-shot VOS, instance-level zero-shot VOS, and one-shot VOS. Experiments demonstrate promising performance in these settings, as well as the potential of MuG in leveraging unlabeled data to further improve the segmentation accuracy.
arXiv commentAccepted to CVPR 2020. Code: https://github.com/carrierlxk/MuG
References in corpus (14)
- Efficient Inference in Fully Connected CRFs with Gaussian Edge Potentials
- End-To-End Memory Networks
- Video Salient Object Detection via Fully Convolutional Networks
- Training Deep Neural Networks on Noisy Labels with Bootstrapping
- RANet: Ranking Attention Network for Fast Video Object Segmentation
- The 2019 DAVIS Challenge on VOS: Unsupervised Multi-Object Segmentation
- Self-supervised Learning for Video Correspondence Flow
- SegFlow: Joint Learning for Video Object Segmentation and Optical Flow
- Efficient Video Object Segmentation via Network Modulation
- Unsupervised Deep Tracking
- FEELVOS: Fast End-to-End Embedding Learning for Video Object Segmentation
- Transitive Invariance for Self-supervised Visual Representation Learning
- Long-term Tracking in the Wild: A Benchmark
- Multigrid Predictive Filter Flow for Unsupervised Learning on Videos