OAMVOS:2nd Report for 5th PVUW MOSE Track

TL;DR

This paper introduces an occlusion- and reappearance-aware extension to DAM4SAM, enhancing long-term robustness of SAM-based dense trackers, especially for small objects, by improving memory control without altering the backbone.

Contribution

It proposes a novel memory control extension for SAM-based trackers that handles occlusion and reappearance more effectively, improving robustness without changing the core backbone.

Findings

Enhanced robustness in occlusion and reappearance scenarios.

Maintains efficiency in easy tracking cases.

Improves long-gap recovery for small objects.

Abstract

SAM-based dense trackers provide strong short-term mask propagation but remain fragile under long occlusion, fast motion, viewpoint change, and distractors. The problem is especially severe for small objects, where a few incorrect memory updates can dominate later predictions. This report presents an occlusion- and reappearance-aware extension of DAM4SAM that improves memory control rather than changing the backbone. The method augments the original SAM3 tracker with four ingredients: a reliability-aware tracking state machine, branch-based recovery, delayed DRM promotion, and a selective policy for native SAM3 memory selection. During stable tracking, the model follows the original single-path propagation process. Once confidence drops, the tracker enters an ambiguous or recovery mode, maintains a small set of candidate branches, and commits memory only after a branch is reconfirmed. For small-object disappearance and reappearance, native memory selection is temporarily bypassed so older anchors remain accessible. In addition, the first conditioning frame is explicitly preserved, and the conditioning-memory budget is moderately enlarged to improve long-gap recovery. The resulting design keeps DAM4SAM efficient in easy cases while improving robustness in sequences dominated by occlusion and reappearance.