Video Object Segmentation with Re-identification

Abstract

Conventional video segmentation methods often rely on temporal continuity to propagate masks. Such an assumption suffers from issues like drifting and inability to handle large displacement. To overcome these issues, we formulate an effective mechanism to prevent the target from being lost via adaptive object re-identification. Specifically, our Video Object Segmentation with Re-identification (VS-ReID) model includes a mask propagation module and a ReID module. The former module produces an initial probability map by flow warping while the latter module retrieves missing instances by adaptive matching. With these two modules iteratively applied, our VS-ReID records a global mean (Region Jaccard and Boundary F measure) of 0.699, the best performance in 2017 DAVIS Challenge.

Published in CVPR 2017 Workshop, DAVIS Challenge on Video Object Segmentation 2017 (Winning Entry)

References in corpus (2)

• Semantic Image Segmentation via Deep Parsing Network
• Lucid Data Dreaming for Video Object Segmentation

Cited by in corpus (12)

• A cow structural model for video analytics of cow health
• Lucid Data Dreaming for Video Object Segmentation
• Anchor Diffusion for Unsupervised Video Object Segmentation
• MHP-VOS: Multiple Hypotheses Propagation for Video Object Segmentation
• U-Net Based Multi-instance Video Object Segmentation
• KinePose: A temporally optimized inverse kinematics technique for 6DOF human pose estimation with biomechanical constraints
• OVSNet : Towards One-Pass Real-Time Video Object Segmentation
• LSMVOS: Long-Short-Term Similarity Matching for Video Object
• LIP: Learning Instance Propagation for Video Object Segmentation
• An Empirical Study of Propagation-based Methods for Video Object Segmentation
• Design Pseudo Ground Truth with Motion Cue for Unsupervised Video Object Segmentation
• Automatic Video Object Segmentation via Motion-Appearance-Stream Fusion and Instance-aware Segmentation