Locate Core, Refine Path: A Training-Free Closed-Loop Paradigm for Referring Video Object Segmentation.

@inproceedings{yu-2026-locate, author = {Jizhe Yu and Hao Zhang and Xiya Bu and Yuhang Duan and Xiaoshuai Wu and Yu Liu}, editor = {Stevan Rudinac and Zi Huang and Marcel Worring and Lucia Vadicamo and Xirong Li and Pascal Mettes and Lizi Liao}, title = {{Locate Core, Refine Path: A Training-Free Closed-Loop Paradigm for Referring Video Object Segmentation}}, booktitle = {{Proceedings of the 2026 International Conference on Multimedia Retrieval, ICMR 2026, Amsterdam, The Netherlands, June 16-19, 2026}}, pages = {1568--1572}, publisher = {ACM}, year = {2026}, url = {https://doi.org/10.1145/3805622.3810836}, doi = {https://doi.org/10.1145/3805622.3810836} }