• Home
  • Search
  • Mask-Guided Self-Supervised Video Object Segmentation.
  • Cite Icon1
  • https://doi.org/10.1109/tpami.2026.3692914Copy DOI Icon

Mask-Guided Self-Supervised Video Object Segmentation.

  • Abstract
  • Literature Map
  • Citations
  • Similar Papers
Abstract

This work aims to learn video object segmentation (mask propagation) in a self-supervised manner. We develop a unified framework which simultaneously models cross-frame dense correspondence for locally discriminative feature learning and embeds object-level context for target-mask decoding. As a result, it is able to directly learn to perform mask-guided sequential segmentation from unlabeled videos, in contrast to previous efforts usually relying on an oblique solution - cheaply "copying" labels according to pixel-wise correlations. Concretely, our algorithm alternates between i) clustering video pixels for creating pseudo segmentation labels ex nihilo; and ii) utilizing the pseudo labels to learn mask encoding and decoding for VOS. Unsupervised correspondence learning is further incorporated into this self-taught, mask embedding scheme, so as to ensure the generic nature of the learnt representation and avoid cluster degeneracy. Building on the above design, we further lift the offline clustering process to an online version, which streamlines the integration of mask embedding and dense correspondence modeling, thus resulting in a more cohesive online learning framework. Consequently, this facilitates more effective generation of pseudo labels without compromising the training speed. Additionally, we introduce a semantic centroids pool, a repository of content-aware visual representations across the entire video data. Managed through multi-frame clustering centroids, it is able to increase the precision and reliability of pseudo segmentation labels. Our improved algorithm sets state-of-the-arts on three standard benchmarks (ie, DAVIS$_{17}$, YouTube-VOS, and VIP), narrowing the gap between self- and fully-supervised VOS. Code is available at Mask-VOS.

Similar Papers
  • Research Article
  • Citations18

Distribution-Guided Hierarchical Calibration Contrastive Network for Unsupervised Person Re-Identification

  • Aug 01, 2024
  • IEEE Transactions on Circuits and Systems for Video Technology
  • Yongxi Li +4
  • Conference Article
  • Citations3

Unsupervised Embedding and Association Network for Multi-Object Tracking

  • Jul 01, 2022
  • Yu-Lei Li
  • PDF
  • Research Article
  • Citations1

Pseudo-label self-training model for transfer learning algorithm

  • Jun 01, 2023
  • Journal of Physics Conference Series
  • Zijie Chen +1
  • Book Chapter
  • Citations106

Two-Phase Pseudo Label Densification for Self-training Based Domain Adaptation

  • Jan 01, 2020
  • Inkyu Shin +3
  • Research Article
  • Citations8

Unsupervised self-training correction learning for 2D image-based 3D model retrieval

  • May 08, 2023
  • Information Processing & Management
  • Yaqian Zhou +5
  • Research Article
  • Citations33

Mask-Aware Pseudo Label Denoising for Unsupervised Vehicle Re-Identification

  • Apr 01, 2023
  • IEEE Transactions on Intelligent Transportation Systems
  • Zefeng Lu +3
  • Book Chapter
  • Citations4

Automatic Semi-supervised Left Atrial Segmentation Using Deep-Supervision 3DResUnet with Pseudo Labeling Approach for LAScarQS 2022 Challenge

  • Jan 01, 2023
  • Moona Mazher +3
  • Research Article

Pseudo Label Learning for Partial Point Cloud Registration.

  • Dec 01, 2025
  • IEEE transactions on visualization and computer graphics
  • Wenping Ma +6
  • Research Article
  • Citations56

Selecting feature subset with sparsity and low redundancy for unsupervised learning

  • Jun 16, 2015
  • Knowledge-Based Systems
  • Jiuqi Han +2
  • Conference Article
  • Citations171

Leveraging Auxiliary Tasks with Affinity Learning for Weakly Supervised Semantic Segmentation

  • Oct 01, 2021
  • Lian Xu +5
  • Conference Article
  • Citations116

Background-Aware Pooling and Noise-Aware Loss for Weakly-Supervised Semantic Segmentation

  • Jun 01, 2021
  • Youngmin Oh +2
  • Conference Article
  • Citations20

Self-Taught Metric Learning without Labels

  • Jun 01, 2022
  • Sungyeon Kim +3
  • Conference Article
  • Citations4

Noise-Tolerant Learning with Silhouette Coefficient for Unsupervised Person Re-Identification

  • Jul 18, 2022
  • Yu Feng +5
  • Research Article
  • Citations3

A novel class restriction loss for unsupervised domain adaptation

  • Jul 20, 2021
  • Neurocomputing
  • Qi He +3
  • Conference Article
  • Citations96

Co-training with High-Confidence Pseudo Labels for Semi-supervised Medical Image Segmentation

  • Aug 01, 2023
  • Zhiqiang Shen +5
Cactus Communications logo

Copyright 2026 Cactus Communications. All rights reserved.