Visual Token Compression

Ask about this topic

Overview

Vision encoders represent images and video as patch or frame tokens, and long token sequences can make multimodal models costly to run, especially during prefill and memory-intensive video reasoning. Methods include training-free pruning and merging, learned resamplers, quantization and codecs, and adaptive token budgets guided by text queries, spatial structure, or temporal redundancy. They are evaluated on tasks such as visual question answering, video understanding, OCR, segmentation, retrieval, generation, and robot control. The central tension is that fewer tokens do not always mean faster end-to-end inference, and aggressive compression can discard details needed for a particular question or task.

Latest papers

CardsList

271 papers46 papers in the last four weeks, up 229% on the four weeks before. 0.3% of all new papers.

  1. Geometry-Guided 3D Visual Token Pruning for Video-Language Models

    Apr 20, 2026Han Li, Zehao Huang, Jiahui Fu +2Visual Token PruningVisual Token Compression

    Beihang University · Zhongguancun Academy

  2. EvoComp: Learning Visual Token Compression for Multimodal Large Language Models via Semantic-Guided Evolutionary Labeling

    Apr 18, 2026Jiafei Song, Fengwei Zhou, Jin Qu +7Efficient Multimodal InferenceMultimodal Large Language Models

    OPPO CTG

  3. Why Training-Free Token Reduction Collapses: The Inherent Instability of Pairwise Scoring Signals

    Apr 17, 2026Yang ShanglinVisual Token PruningVisual Token Compression

  4. StreamCacheVGGT: Streaming Visual Geometry Transformers with Robust Scoring and Hybrid Cache Compression

    Apr 16, 2026Xuanyi Liu, Chunan Yu, Deyi Ji +63D ViTsStreaming 3D Reconstruction

    KOKONI 3D, Moxin Technology · Nanjing University of Science and Technology · University of Science and Technology of China +4

  5. Revisiting Token Compression for Accelerating ViT-based Sparse Multi-View 3D Object Detectors

    Apr 16, 2026Mingqian Ji, Shanshan Zhang, Jian YangMulti-View 3D Object DetectionEfficient ViTs

    PCA Lab, School of Computer Science and Engineering, Nanjing University of Science and Technology · PCA Lab, School of Intelligence Science and Technology, Nanjing University

  6. UHR-BAT: Budget-Aware Token Compression Vision-Language model for Ultra-High-Resolution Remote Sensing

    Apr 15, 2026Yunkai Dang, Minxin Dai, Yuekun Yang +4Visual Token CompressionRemote Sensing Image Understanding

    School of Artificial Intelligence Science and Technology, Nanjing University · School of Electronic Science and Engineering, Nanjing University · School of Physics, Nanjing University

  7. MaMe & MaRe: Matrix-Based Token Merging and Restoration for Efficient Visual Perception and Synthesis

    Apr 15, 2026Simin Huo, Ning LiVisual Token MergingVisual Token Compression

    Shanghai Jiao Tong University, , Shanghai, 200240, Shanghai, China

  8. POINTS-Long: Adaptive Dual-Mode Visual Reasoning in MLLMs

    Apr 13, 2026Haicheng Wang, Yuan Liu, Yikun Liu +9Multimodal Large Language ModelsLong-Context VLMs

    SAI, Shanghai Jiao Tong University, China · WeChat AI, Tencent, China

  9. Semantic-Geometric Dual Compression: Training-Free Visual Token Reduction for Ultra-High-Resolution Remote Sensing Understanding

    Apr 13, 2026Yueying Li, Fengxiang Wang, Yan Li +3Visual Token CompressionEfficient VLM Inference

    College of Computer Science and Technology, National University of Defense Technology, Changsha 410073, China · School of Computer Science, Beijing University of Posts and Telecommunications, Beijing 100876, China

  10. Visual Late Chunking: An Empirical Study of Contextual Chunking for Efficient Visual Document Retrieval

    Apr 11, 2026Yibo Yan, Mingdong Ou, Yi Cao +5Visual Document RetrievalMulti-Vector Retrieval

    Hong Kong University of Science and Technology (Guangzhou) · Alibaba Cloud Computing · Hong Kong University of Science and Technology

  11. UIPress: Bringing Optical Token Compression to UI-to-Code Generation

    Apr 10, 2026Dasen Dai, Shuoqi Li, Ronghao Chen +3Token CompressionVisual Token Compression

    The Chinese University of Hong Kong China · Peking University China · University of Chinese Academy of Sciences China +2

  12. Do Vision Language Models Need to Process Image Tokens?

    Apr 10, 2026Sambit Ghosh, R. Venkatesh Babu, Chirag AgarwalVision-Language ModelsVLM Reasoning

    IBM · Indian Institute of Science · University of Virginia

  13. EPIR: An Efficient Patch Tokenization, Integration and Representation Framework for Micro-expression Recognition

    Apr 9, 2026Junbo Wang, Liangyu Fu, Yuke Li +3Micro-Expression RecognitionFacial Expression Recognition

    School of Software, Northwestern Polytechnical University, Xi’an, 710129, China · School of Computer Science, Northwestern Polytechnical University, Xi’an, 710129, China · School of Computer Science and Technology, Xi’an Jiaotong University, Xi’an, 710049, China +1

  14. TC-AE: Unlocking Token Capacity for Deep Compression Autoencoders

    Apr 8, 2026Teng Li, Ziyuan Huang, Cong Chen +5AutoencodersVisual Token Compression

    Inclusion AI, Ant Group · HKUST · ZJU +2

  15. Hierarchical SVG Tokenization: Learning Compact Visual Programs for Scalable Vector Graphics Modeling

    Apr 6, 2026Ximing Xing, Ziteng Xue, Zhenxi Li +8Vector Graphics GenerationVisual Tokenization

    Tencent · Visual Computing Group, Beijing

  16. RCP: Representation Consistency Pruner for Mitigating Distribution Shift in Large Vision-Language Models

    Apr 4, 2026Jianwei Zhang, Chaoning Zhang, Sihan Cao +7Efficient VLM InferenceLarge Vision-Language Models

    School of Computer Science and Engineering, University of Electronic Science and Technology of China, Qingshuihe Campus, No. 2006 Xiyuan Avenue, High-Tech Zone (West), Chengdu 611731, China · Machine Learning Department, Mohamed bin Zayed University of Artificial Intelligence, Building 1B, Masdar City, Abu Dhabi, United Arab Emirates · School of Robotics and Advanced Manufacture, Harbin Institute of Technology, Shenzhen, University Town of Shenzhen, Nanshan District, Shenzhen 518055, China +1

  17. KiToke: Distribution-Aware Token Compression for Video Large Language Models

    Apr 3, 2026Haifeng Huang, Biyin Xu, Yang LiVideo Token CompressionVisual Token Compression

    Department of Computer Science, Iowa State University.

  18. PixelPrune: Pixel-Level Adaptive Visual Token Reduction via Predictive Coding

    Apr 1, 2026Nan Wang, Zhiwei Jin, Chen Chen +1Efficient VLM InferenceVisual Token Pruning

    OPPO AI Center

  19. MacTok: Robust Continuous Tokenization for Image Generation

    Mar 31, 2026Hengyu Zeng, Xin Gao, Guanghao Li +5Image TokenizationVisual Token Compression

    Fudan University

  20. Towards Efficient Large Vision-Language Models: A Comprehensive Survey on Inference Strategies

    Mar 30, 2026Surendra Pathak, Bo HanEfficient VLM InferenceVisual Token Compression

    Computer Science Department, George Mason University, Fairfax, VA - 22030, USA

  21. Semantic-Aware Prefix Learning for Token-Efficient Image Generation

    Mar 26, 2026Qingfeng Li, Haoxian Zhang, Xu He +4Visual TokenizationVisual Token Compression

    Institute of Automation, Chinese Academy of Sciences · Kling Team, Kuaishou Technology · Tsinghua University

  22. Photon: Speedup Volume Understanding with Efficient Multimodal Large Language Models

    Mar 26, 2026Chengyu Fang, Heng Guo, Zheng Jiang +3Medical VLMs3D VLMs

    Tsinghua University · DAMO Academy, Alibaba Group · Hupan Lab +1

  23. Rethinking Token Reduction for Large Vision-Language Models

    Mar 23, 2026Yi Wang, Haofei Zhang, Qihan Huang +7Efficient VLM InferenceVisual Token Compression

    College of Computer Science and Technology, Zhejiang University · State Key Laboratory of Blockchain and Data Security, Zhejiang University · Hangzhou High-Tech Zone (Binjiang) Institute of Blockchain and Data Security +3

  24. QMoP: Query Guided Mixture-of-Projector for Efficient Visual Token Compression

    Mar 22, 2026Zhongyang Li, Yaqian Li, Faming Fang +6Multimodal Token CompressionVisual Token Compression

    East China Normal University · Li Auto Inc.

  25. CurveStream: Boosting Streaming Video Understanding in MLLMs via Curvature-Aware Hierarchical Visual Memory Management

    Mar 20, 2026Chao Wang, Xudong Tan, Jianjian Cao +2Streaming Video UnderstandingMemory-Augmented Video Understanding

    College of Future Information Technology, Fudan University, Shanghai, China · Shanghai Innovation Institute, Shanghai, China

  26. Multimodal Model for Computational Pathology:Representation Learning and Image Compression

    Mar 19, 2026Peihang Wu, Zehong Chen, Lijian XuComputational PathologyWhole-Slide Image Analysis

    Shenzhen University of Advanced Technology

  27. STAC: Plug-and-Play Spatio-Temporal Aware Cache Compression for Streaming 3D Reconstruction

    Mar 18, 2026Runze Wang, Yuxuan Song, Youcheng Cai +1Streaming 3D ReconstructionKV-Cache Compression

    University of Science and Technology of China

  28. InViC: Intent-aware Visual Cues for Medical Visual Question Answering

    Mar 17, 2026Zhisong Wang, Ziyang Chen, Zanting Ye +3Medical VQAMedical VLMs

    National Engineering Laboratory for Integrated Aero-Space-Ground-Ocean Big Data Application Technology, School of Computer Science and Engineering, Northwestern Polytechnical University, Xi’an 710072, China · Southern Medical University · Westlake University

  29. Question-guided Visual Compression with Memory Feedback for Long-Term Video Understanding

    Mar 16, 2026Sosuke Yamao, Natsuki Miyahara, Yuankai Qi +1Long-Video UnderstandingTemporal Video Understanding

    Fujitsu Research · Macquarie University

  30. Balancing Saliency and Coverage: Semantic Prominence-Aware Budgeting for Visual Token Compression in VLMs

    Mar 16, 2026Jaehoon Lee, Mingi Jung, Soohyuk Jang +3Efficient VLM InferenceVisual Token Pruning

    Interdisciplinary Program in AI, Seoul National University · Department of Electrical and Computer Engineering, Seoul National University · Department of Artificial Intelligence, Chung-Ang University