Thinking With Bounding Boxes: Enhancing Spatio-Temporal Video Grounding via Reinforcement Fine-Tuning

• URL: http://arxiv.org/abs/2511.21375v1
• Date: Wed, 26 Nov 2025 13:21:15 GMT
• Title: Thinking With Bounding Boxes: Enhancing Spatio-Temporal Video Grounding via Reinforcement Fine-Tuning
• Authors: Xin Gu, Haoji Zhang, Qihang Fan, Jingxuan Niu, Zhipeng Zhang, Libo Zhang, Guang Chen, Fan Chen, Longyin Wen, Sijie Zhu,
• Abstract summary: multimodal large language models (LMs) underperform on STVG due to misaligned training objectives and weak fine-grained fine-word alignment in standard visual encoders.<n>We propose STVG-o1, the first framework that enables off-the-shelf MLLMs to achieve state-of-the-temporal STVG performance without architectural modifications.
• Score: 41.30900315121155
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Spatio-temporal video grounding (STVG) requires localizing a target object in untrimmed videos both temporally and spatially from natural language descriptions. Despite their strong language understanding, multimodal large language models (MLLMs) underperform on STVG due to misaligned training objectives and weak fine-grained region-word alignment in standard visual encoders. To address this, we propose STVG-o1, the first framework that enables off-the-shelf MLLMs to achieve state-of-the-art STVG performance without any architectural modifications. Our method introduces a bounding-box chain-of-thought mechanism that explicitly reasons about spatio-temporal locations in an intermediate step before producing the final prediction. We further design a multi-dimensional reinforcement reward function consisting of format, consistency, temporal, spatial, and think rewards, which provides geometry-aware supervision through reinforcement fine-tuning. Evaluated on HCSTVG-v1/v2 and VidSTG, STVG-o1 sets new state-of-the-art results on HCSTVG, outperforming the best task-specific method by 7.3\% m\_tIoU on HCSTVG-v1, matching specialized models on VidSTG, and surpassing all existing MLLM-based approaches by large margins. It also demonstrates strong open-vocabulary generalization across datasets, establishing MLLMs as viable and powerful backbones for precise spatio-temporal grounding. Our code and models will be released.

Related papers

• STVG-R1: Incentivizing Instance-Level Reasoning and Grounding in Videos via Reinforcement Learning [65.36458157092207]
  In vision-language models (VLMs), misalignment between textual descriptions and visual coordinates often induces hallucinations.<n>We propose a novel visual prompting paradigm that avoids the difficult problem of aligning coordinates across modalities.<n>We introduce STVG-R1, the first reinforcement learning framework for STVG, which employs a task-driven reward to jointly optimize temporal accuracy, spatial consistency, and structural format regularization.
  arXiv  Detail & Related papers  (2026-02-12T08:53:32Z)
• VOST-SGG: VLM-Aided One-Stage Spatio-Temporal Scene Graph Generation [18.15310805625469]
  VOST-SGG is a VLM-aided one-stage ST-SGG framework that integrates the common sense reasoning capabilities of vision-language models.<n>We propose a multi-modal feature bank that fuses visual, textual, and spatial cues for improved predicate classification.<n>Our approach achieves state-of-the-art performance, validating the effectiveness of integrating VLM-aided semantic priors and multi-modal features for ST-SGG.
  arXiv  Detail & Related papers  (2025-12-05T08:34:06Z)
• Unleashing the Potential of Multimodal LLMs for Zero-Shot Spatio-Temporal Video Grounding [47.400649582392255]
  We use large language models (MLLMs) to explore a zero-shot solution in STVG.<n>We propose a MLLM-based zero-shot framework for STVG, which includes novel temporal-augmented assembling strategies.
  arXiv  Detail & Related papers  (2025-09-18T17:35:50Z)
• MLVTG: Mamba-Based Feature Alignment and LLM-Driven Purification for Multi-Modal Video Temporal Grounding [13.025856914576673]
  Video Temporal Grounding aims to localize video clips corresponding to natural language queries.<n>Existing Transformer-based methods often suffer from redundant attention and suboptimal multi-modal alignment.<n>We propose MLVTG, a novel framework that integrates two key modules: MambaAligner and LLMRefiner.
  arXiv  Detail & Related papers  (2025-06-10T07:20:12Z)
• VideoMolmo: Spatio-Temporal Grounding Meets Pointing [66.19964563104385]
  VideoMolmo is a model tailored for fine-grained pointing of video sequences.<n>A novel temporal mask fusion employs SAM2 for bidirectional point propagation.<n>To evaluate the generalization of VideoMolmo, we introduce VPoMolS-temporal, a challenging out-of-distribution benchmark spanning five real-world scenarios.
  arXiv  Detail & Related papers  (2025-06-05T17:59:29Z)
• Enhancing Video-Language Representations with Structural Spatio-Temporal Alignment [130.15775113897553]
  Finsta is a fine-grained structural-temporal alignment learning method. It consistently improves the existing 13 strong-tuning video-language models.
  arXiv  Detail & Related papers  (2024-06-27T15:23:36Z)
• Video-GroundingDINO: Towards Open-Vocabulary Spatio-Temporal Video Grounding [108.79026216923984]
  Video grounding aims to localize a-temporal section in a video corresponding to an input text query. This paper addresses a critical limitation in current video grounding methodologies by introducing an Open-Vocabulary Spatio-Temporal Video Grounding task.
  arXiv  Detail & Related papers  (2023-12-31T13:53:37Z)
• Embracing Consistency: A One-Stage Approach for Spatio-Temporal Video Grounding [35.73830796500975]
  We present an end-to-end one-stage framework, termed Spatio-Temporal Consistency-Aware Transformer (STCAT) To generate the above template under sufficient video- perception, an encoder-decoder architecture is proposed for effective global context modeling. Our method outperforms previous state-of-the-arts with clear margins on two challenging video benchmarks.
  arXiv  Detail & Related papers  (2022-09-27T11:13:04Z)
• Spatio-Temporal Ranked-Attention Networks for Video Captioning [34.05025890230047]
  We propose a model that combines spatial and temporal attention to videos in two different orders. We provide experiments on two benchmark datasets: MSVD and MSR-VTT. Our results demonstrate the synergy between the ST and TS modules, outperforming recent state-of-the-art methods.
  arXiv  Detail & Related papers  (2020-01-17T01:00:45Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.