1st Place Solution for 5th LSVOS Challenge: Referring Video Object Segmentation
This work addresses video object segmentation for applications like video editing and autonomous systems, but it is incremental as it builds on existing transformer-based models.
The paper tackles the Referring Video Object Segmentation (RVOS) task by integrating leading models and proposing a Two-Stage Multi-Model Fusion strategy to improve mask consistency and quality, achieving 75.7% J&F on the Ref-Youtube-VOS validation set and 70% J&F on the test set, securing first place in the 5th LSVOS Challenge.
The recent transformer-based models have dominated the Referring Video Object Segmentation (RVOS) task due to the superior performance. Most prior works adopt unified DETR framework to generate segmentation masks in query-to-instance manner. In this work, we integrate strengths of that leading RVOS models to build up an effective paradigm. We first obtain binary mask sequences from the RVOS models. To improve the consistency and quality of masks, we propose Two-Stage Multi-Model Fusion strategy. Each stage rationally ensembles RVOS models based on framework design as well as training strategy, and leverages different video object segmentation (VOS) models to enhance mask coherence by object propagation mechanism. Our method achieves 75.7% J&F on Ref-Youtube-VOS validation set and 70% J&F on test set, which ranks 1st place on 5th Large-scale Video Object Segmentation Challenge (ICCV 2023) track 3. Code is available at https://github.com/RobertLuo1/iccv2023_RVOS_Challenge.