不训练模型，通过双路记忆分离实现高效视频目标分割
SAM3Dual: A 3rd Place Solution to the MOSEv2 Track, 8th LSVOS Challenge
- 将时序记忆分为短期与长期分支，分别处理近期和历史信息
- 推理时融合双路输出，获得64.37的官方J&F得分，排名第三
- 完全冻结预训练参数，无需微调或在线优化，适合快速部署
我们提出SAM3Dual，作为在ECCV 2026第八届大规模视频目标分割（LSVOS）挑战赛中MOSEv2赛道的第三名解决方案。SAM3Dual是预训练SAM 3的无训练推理扩展，明确将时序记忆拆分为短期分支（处理近期观测）和长期分支（处理间隔采样的历史表示）。两条路径的响应通过确定性的序列相对融合策略结合，并由前一帧的目标置信度保守调制。所有预训练SAM 3参数保持冻结，无需任务特定训练、微调、测试时训练或在线参数优化。完整系统在官方评测中取得64.37的J&F得分，在MOSEv2赛道中排名第三。该结果凸显了仅在推理阶段重构时序记忆即可实现具有竞争力的长时视频目标分割性能，同时保留预训练模型能力。
原文摘要 · Abstract (English)
We present SAM3Dual, our third-place solution to the MOSEv2 track of the 8th Large-scale Video Object Segmentation (LSVOS) Challenge at ECCV 2026. SAM3Dual is a training-free inference extension of pretrained SAM 3 that explicitly separates temporal memory into a short-term branch for recent observations and a long-term branch for interval-sampled historical representations. The two memory responses are combined using a deterministic sequence-relative fusion schedule and conservatively modulated by the previous-frame object confidence. All pretrained SAM 3 parameters remain frozen, requiring no task-specific training, fine-tuning, test-time training, or online parameter optimization. The complete system achieved an official J&F score of 64.37 and ranked third in the MOSEv2 track. This result highlights the potential of reorganizing temporal memory entirely at inference time to obtain competitive long-term VOS performance while preserving the pretrained model.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。