mlsanityData to 7 Oct 2026

Dynamic Resolution Routing for Efficient Egocentric Grounding

arXiv 2608.01638 ↗First posted 3 Aug 2026cs.CVcited by 0

On the map: Efficient LLM inference and serving → Visual token pruning in multimodal models

Abstract (arXiv)

Egocentric visual grounding requires high-resolution inputs to localize small objects. However, scaling Multimodal Large Language Models to this domain is constrained by the excessive cost of visual token processing. We identify that current efficient strategies based on token reduction are unreliable for selecting object-centric spatial evidence. To overcome this, we propose SmartRes, a framework that performs efficiency optimization in the pixel space via dynamic resolution routing. SmartRes first encodes a low-resolution view for global context and uses a lightweight router to activate high-resolution patches in object-centric regions and constructs an order-preserving visual sequence. To further enable robust routing under severe foreground-background imbalance, we introduce a margin-regularized routing objective that increases foreground-background logit separation and improves foreground recall. Experiments on Ego4D and EgoIntention show that SmartRes reduces visual tokens by up to 67% while retaining 86.4% of full-resolution performance, and achieves up to 1.66X faster inference than state-of-the-art token reduction methods with higher accuracy. Furthermore, strong performance on small object grounding indicates the effectiveness of SmartRes towards egocentric applications. Code will be publicly available.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Visual token pruning in multimodal models, one of 12 directions in Efficient LLM inference and serving. Builds on 38 of its 46 references that are in the map. First posted on arXiv 3 Aug 2026.

Builds on

38 of 46 references.

Closest ancestors · 6

Same direction, newest first

  1. 2025Don't Just Chase "Highlighted Tokens" in MLLMs: Revisiting Visual Holistic Context Retention3 Oct
  2. 2025VScan: Rethinking Visual Token Reduction for Efficient Large Vision-Language Models28 May
  3. 2025Local Information Matters: Inference Acceleration For Grounded Conversation Generation Models Through Adaptive Local-Aware Token Pruning31 Mar
  4. 2024Dynamic-LLaVA: Efficient Multimodal Large Language Models via Dynamic Vision-language Context Sparsification1 Dec
  5. 2024ATP-LLaVA: Adaptive Token Pruning for Large Vision Language Models30 Nov
  6. 2024An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models11 Mar

Other references · 32

  1. 2025Qwen2.5-VL Technical Reportcited by 5,418
  2. 2024Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolutioncited by 4,151
  3. 2025Qwen3-VL Technical Reportcited by 2,534
  4. 2016V-Net: Fully Convolutional Neural Networks for Volumetric Medical Image Segmentationcited by 2,284
  5. 2023Qwen-VL: A Versatile Vision-Language Model for Understanding, Localization, Text Reading, and Beyondcited by 2,179
  6. 2013Estimating or Propagating Gradients Through Stochastic Neurons for Conditional Computationcited by 2,031
  7. 2021Ego4D: Around the World in 3,000 Hours of Egocentric Videocited by 1,530
  8. 2019Deep High-Resolution Representation Learning for Human Pose Estimationcited by 1,339
  9. 2019Class-Balanced Loss Based on Effective Number of Samplescited by 1,239
  10. 2016Modeling Context in Referring Expressionscited by 1,054
  11. 2022Token Merging: Your ViT But Fastercited by 995
  12. 2015Generation and Comprehension of Unambiguous Object Descriptionscited by 897
20 more
  1. 2023Vision Transformers Need Registerscited by 844
  2. 2023V*: Guided Visual Search as a Core Mechanism in Multimodal LLMscited by 459
  3. 2024SparseVLM: Visual Token Sparsification for Efficient Vision-Language Model Inferencecited by 378
  4. 2024VisionZip: Longer is Better but Not Necessary in Vision Language Modelscited by 298
  5. 2024PyramidDrop: Accelerating Your Large Vision-Language Models via Pyramid Visual Redundancy Reductioncited by 283
  6. 2016Modeling Context Between Objects for Referring Expression Understandingcited by 250
  7. 2017Tversky loss function for image segmentation using 3D fully convolutional deep networkscited by 228
  8. 2025MLLMs Know Where to Look: Training-free Perception of Small Visual Details with Multimodal LLMscited by 172
  9. 2024TokenPacker: Efficient Visual Projector for Multimodal LLMcited by 170
  10. 2023PACO: Parts and Attributes of Common Objectscited by 146
  11. 2023LLaVA-Grounding: Grounded Visual Chat with Large Multimodal Modelscited by 113
  12. 2025VisionThink: Smart and Efficient Vision Language Model via Reinforcement Learningcited by 41
  13. 2025Scaling Vision Pre-Training to 4K Resolutioncited by 26
  14. 2023RefEgo: Referring Expression Comprehension Dataset from First-Person Perception of Ego4Dcited by 21
  15. 2025Patch-as-Decodable-Token: Towards Unified Multi-Modal Vision Tasks in MLLMscited by 20
  16. 2025AdaptVision: Efficient Vision-Language Models via Adaptive Visual Acquisitioncited by 19
  17. 2024HiRes-LLaVA: Restoring Fragmentation Input in High-Resolution Large Vision-Language Modelscited by 16
  18. 2025Visual Intention Grounding for Egocentric Assistantscited by 11
  19. 2026FocusUI: Efficient UI Grounding via Position-Preserving Visual Token Selectioncited by 9
  20. 2025Zoomer: Adaptive Image Focus Optimization for Black-box MLLMcited by 5

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026TempoGround: State-Aware Streaming Visual Grounding with Vision-Language Models
  2. 2026Focus When Necessary: Adaptive Routing and Collaborative Grounding for Training-Free Visual Grounding
  3. 2025From Objects to Anywhere: A Holistic Benchmark for Multi-level Visual Grounding in 3D Scenes
  4. 2026Beyond Language: Grounding Referring Expressions with Hand Pointing in Egocentric Vision
  5. 2026Beyond Static Cropping: Layer-Adaptive Visual Localization and Decoding Enhancement
  6. 2026Groundbench: Multi-Resolution Polygon Grounding Exposes the Geometry Gap in Vision-Language Models
  7. 2026UniGround: Universal 3D Visual Grounding via Training-Free Scene Parsing
  8. 2026CoEvolve: Construct-to-Edit Visual Grounding with Bidirectional State Refinement