Video Spatial Reasoning with Object-Centric 3D Rollout
This addresses the challenge of query-locked reasoning in video spatial understanding for AI systems, representing a strong domain-specific advancement.
The paper tackled the problem of enabling robust video spatial reasoning in Multi-modal Large Language Models (MLLMs) by proposing Object-Centric 3D Rollout (OCR), which achieved state-of-the-art performance with 47.5% accuracy on VSI-Bench using a 3B-parameter model, outperforming larger 7B baselines.
Recent advances in Multi-modal Large Language Models (MLLMs) have showcased remarkable capabilities in vision-language understanding. However, enabling robust video spatial reasoning-the ability to comprehend object locations, orientations, and inter-object relationships in dynamic 3D scenes-remains a key unsolved challenge. Existing approaches primarily rely on spatially grounded supervised fine-tuning or reinforcement learning, yet we observe that such models often exhibit query-locked reasoning, focusing narrowly on objects explicitly mentioned in the prompt while ignoring critical contextual cues. To address this limitation, we propose Object-Centric 3D Rollout (OCR), a novel strategy that introduces structured perturbations to the 3D geometry of selected objects during training. By degrading object-specific visual cues and projecting the altered geometry into 2D space, OCR compels the model to reason holistically across the entire scene. We further design a rollout-based training pipeline that jointly leverages vanilla and region-noisy videos to optimize spatial reasoning trajectories. Experiments demonstrate state-of-the-art performance: our 3B-parameter model achieves 47.5% accuracy on VSI-Bench, outperforming several 7B baselines. Ablations confirm OCR's superiority over prior rollout strategies (e.g., T-GRPO, NoisyRollout).