BoxTuning: Directly Injecting the Object Box for Multimodal Model Fine-Tuning

• URL: http://arxiv.org/abs/2604.11136v1
• Date: Mon, 13 Apr 2026 07:49:31 GMT
• Title: BoxTuning: Directly Injecting the Object Box for Multimodal Model Fine-Tuning
• Abstract summary: BoxTuning injects object spatial-temporal information directly into the visual modality.<n>It reduces the token cost significantly, achieving 87-93% text token reduction in practice.<n>It also preserves full temporal resolution, where the trajectory trails further encode inter-frame motion direction and speed within each.
• Score: 18.859227273698615
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Object-level spatial-temporal understanding is essential for video question answering, yet existing multimodal large language models (MLLMs) encode frames holistically and lack explicit mechanisms for fine-grained object grounding. Recent work addresses this by serializing bounding box coordinates as text tokens, but this text-coordinate paradigm suffers from a fundamental modality mismatch: object information is inherently visual, yet encoding it as text incurs a high token cost that forces aggressive temporal downsampling. We propose BoxTuning, which resolves this mismatch by injecting object spatial-temporal information directly into the visual modality. Colored bounding boxes and trajectory trails are rendered onto video frames as visual prompts, with only a concise color-to-object legend retained as text. This reduces the token cost significantly, achieving 87-93% text token reduction in practice. It also preserves full temporal resolution, where the trajectory trails further encode inter-frame motion direction and speed within each keyframe, recovering fine-grained dynamics that text-coordinate methods are forced to discard. Experimental results on five video QA benchmarks (CLEVRER, Perception Test, STAR, NExT-QA, IntentQA) show that BoxTuning surpasses text-coordinate baselines on spatially oriented tasks and nearly eliminates the accuracy degradation observed on reasoning-centric tasks, establishing visual prompting as a more natural and efficient paradigm for conveying object information to video MLLMs.

Related papers

• ENEAS: Embedding-guided Neural Ensemble for Adaptive Segmentation [36.94429692322632]
  We present ENEAS, a text-promptable method for instance tracking and semantic discovery.<n>We extend the geometrically robust SeC architecture, previously limited to point interactions, with a text-prompting adapter.<n>The verification layer combines high-speed visual embedding matching with conditional VLM refinement, invoking semantic reasoning only for ambiguous candidates.
  arXiv  Detail & Related papers  (2026-09-03T12:25:26Z)
• Persistent Object Narratives for Token-Efficient Video Language Models [10.96152392810444]
  Video large language models (Video-LLMs) have made strong progress in open-ended video understanding.<n>We introduce SlotNarrative, a slot-based interface that organizes a video into persistent object narratives represented by compact object-state tokens.
  arXiv  Detail & Related papers  (2026-08-05T13:59:04Z)
• CoT-Edit: Let CoT Guide Instruction Video Editing [13.600094788792633]
  We propose a plan--guide--edit framework that bridges semantic intent and spatial execution.<n>In our framework, a Chain-of-Thought (CoT)-enhanced multimodal large language model (MLLM) serves as a planner.<n>These spatial priors then guide a box-conditioned mask generator, transforming ambiguous global retrieval into localized, context-aware refinement.
  arXiv  Detail & Related papers  (2026-08-02T09:20:15Z)
• USS: Unified Spatial-Semantic Prompts for Embodied Visual Tracking with Latent Dynamics Learning [18.003030497082623]
  Embodied Visual Tracking (EVT) requires an agent to continuously follow a specified target while actively moving through dynamic environments.<n>We propose a paradigm shift, reframing target indication in EVT from text-only specification to unified spatial-semantic prompting.<n>We introduce Unified Spatial-Semantic Prompts for Embodied Visual Tracking with Latent Dynamics Learning.
  arXiv  Detail & Related papers  (2026-06-24T14:25:05Z)
• Token Reduction via Local and Global Contexts Optimization for Efficient Video Large Language Models [61.11154533305096]
  Video Large Language Models (VLLMs) demonstrate strong video understanding but suffer from inefficiency due to redundant visual tokens.<n>We propose a new perspective that elaborates token textbfAnchors within intra-frame and inter-frame contexts.<n>Our proposed AOT obtains competitive performances across various short- and long-video benchmarks on leading video LLMs.
  arXiv  Detail & Related papers  (2026-03-02T03:06:40Z)
• Rethinking Visual Token Reduction in LVLMs under Cross-modal Misalignment [38.04426918886084]
  Vision-Language Models (LVLMs) encode visual inputs as dense sequences of patch-level tokens to capture fine-grained semantics.<n>Previous efforts have explored visual token reduction either prior to or within the large language models (LLMs)<n>We introduce VisionDrop, a training-free, visual-only pruning framework that selects informative visual tokens based on intra-modal (visual-to-visual) attention.
  arXiv  Detail & Related papers  (2025-06-27T14:55:40Z)
• Caption Anything in Video: Fine-grained Object-centric Captioning via Spatiotemporal Multimodal Prompting [60.58915701973593]
  We present CAT-V (Caption AnyThing in Video), a training-free framework for fine-grained object-centric video captioning.<n>Cat-V integrates three key components: a Segmenter based on SAMI for precise object segmentation across frames, a Temporal Analyzer powered by TRACE-UniVL, and a Captioner using Intern-2.5.<n>Our framework generates detailed, temporally-aware descriptions of objects' attributes, actions, statuses, interactions, and environmental contexts without requiring additional training data.
  arXiv  Detail & Related papers  (2025-04-07T22:35:36Z)
• HiTVideo: Hierarchical Tokenizers for Enhancing Text-to-Video Generation with Autoregressive Large Language Models [63.65066762436074]
  HiTVideo aims to address the potential limitations of existing video tokenizers in text-to-video generation tasks.<n>It utilizes a 3D causal VAE with a multi-layer discrete token framework, encoding video content into hierarchically structured codebooks.
  arXiv  Detail & Related papers  (2025-03-14T15:36:39Z)
• Spatio-temporal Graph Learning on Adaptive Mined Key Frames for High-performance Multi-Object Tracking [5.746443489229576]
  Key Frame Extraction (KFE) module leverages reinforcement learning to adaptively segment videos.<n> Intra-Frame Feature Fusion (IFF) module uses a Graph Convolutional Network (GCN) to facilitate information exchange between the target and surrounding objects.<n>Our proposed tracker achieves impressive results on the MOT17 dataset.
  arXiv  Detail & Related papers  (2025-01-17T11:36:38Z)
• Coarse Correspondences Boost Spatial-Temporal Reasoning in Multimodal Language Model [51.83436609094658]
  We introduce Coarse Correspondences, a simple lightweight method that enhances MLLMs' spatial-temporal reasoning with 2D images as input. Our method uses a lightweight tracking model to identify primary object correspondences between frames in a video or across different image viewpoints. We demonstrate that this simple training-free approach brings substantial gains to GPT4-V/O consistently on four benchmarks.
  arXiv  Detail & Related papers  (2024-08-01T17:57:12Z)
• Rethinking Image-to-Video Adaptation: An Object-centric Perspective [61.833533295978484]
  We propose a novel and efficient image-to-video adaptation strategy from the object-centric perspective. Inspired by human perception, we integrate a proxy task of object discovery into image-to-video transfer learning.
  arXiv  Detail & Related papers  (2024-07-09T13:58:10Z)
• Object-Centric Diffusion for Efficient Video Editing [64.71639719352636]
  Diffusion-based video editing has reached impressive quality. Such solutions typically incur heavy memory and computational costs to generate temporally-coherent frames. We propose modifications that allow significant speed-ups whilst maintaining quality.
  arXiv  Detail & Related papers  (2024-01-11T08:36:15Z)
• Learning Referring Video Object Segmentation from Weak Annotation [78.45828085350936]
  Referring video object segmentation (RVOS) is a task that aims to segment the target object in all video frames based on a sentence describing the object. We propose a new annotation scheme that reduces the annotation effort by 8 times, while providing sufficient supervision for RVOS. Our scheme only requires a mask for the frame where the object first appears and bounding boxes for the rest of the frames.
  arXiv  Detail & Related papers  (2023-08-04T06:50:52Z)
• Learning Spatio-Temporal Transformer for Visual Tracking [108.11680070733598]
  We present a new tracking architecture with an encoder-decoder transformer as the key component. The whole method is end-to-end, does not need any postprocessing steps such as cosine window and bounding box smoothing. The proposed tracker achieves state-of-the-art performance on five challenging short-term and long-term benchmarks, while running real-time speed, being 6x faster than Siam R-CNN.
  arXiv  Detail & Related papers  (2021-03-31T15:19:19Z)

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.