Bridging Time and Space: Decoupled Spatio-Temporal Alignment for Video Grounding
2026-04-09 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors study the problem of finding where and when objects appear in videos based on language descriptions, which is hard because it mixes time and space information. They propose Bridge-STG, a method that separates locating objects in time and space to make it easier, while using special techniques to keep the meaning connected between these steps. Their method reduces irrelevant information and improves accuracy compared to similar models. They show that Bridge-STG does better on standard tests and works well for different video tasks.
Spatio-Temporal Video GroundingMultimodal Large Language Models (MLLMs)Temporal localizationSpatial localizationSemantic bridgingVisual token redundancyTransformer decoderVideo understandingMulti-task learning
Authors
Xuezhen Tu, Jingyu Wu, Fangyu Kang, Qingpeng Nong, Kaijin Zhang, Chaoyue Niu, Fan Wu
Abstract
Spatio-Temporal Video Grounding requires jointly localizing target objects across both temporal and spatial dimensions based on natural language queries, posing fundamental challenges for existing Multimodal Large Language Models (MLLMs). We identify two core challenges: \textit{entangled spatio-temporal alignment}, arising from coupling two heterogeneous sub-tasks within the same autoregressive output space, and \textit{dual-domain visual token redundancy}, where target objects exhibit simultaneous temporal and spatial sparsity, rendering the overwhelming majority of visual tokens irrelevant to the grounding query. To address these, we propose \textbf{Bridge-STG}, an end-to-end framework that decouples temporal and spatial localization while maintaining semantic coherence. While decoupling is the natural solution to this entanglement, it risks creating a semantic gap between the temporal MLLM and the spatial decoder. Bridge-STG resolves this through two pivotal designs: the \textbf{Spatio-Temporal Semantic Bridging (STSB)} mechanism with Explicit Temporal Alignment (ETA) distills the MLLM's temporal reasoning context into enriched bridging queries as a robust semantic interface; and the \textbf{Query-Guided Spatial Localization (QGSL)} module leverages these queries to drive a purpose-built spatial decoder with multi-layer interactive queries and positive/negative frame sampling, jointly eliminating dual-domain visual token redundancy. Extensive experiments across multiple benchmarks demonstrate that Bridge-STG achieves state-of-the-art performance among MLLM-based methods. Bridge-STG improves average m\_vIoU from $26.4$ to $34.3$ on VidSTG and demonstrates strong cross-task transfer across various fine-grained video understanding tasks under a unified multi-task training regime.