{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,4]],"date-time":"2026-05-04T03:29:55Z","timestamp":1777865395716,"version":"3.51.4"},"reference-count":84,"publisher":"IEEE","license":[{"start":{"date-parts":[[2025,10,19]],"date-time":"2025-10-19T00:00:00Z","timestamp":1760832000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2025,10,19]],"date-time":"2025-10-19T00:00:00Z","timestamp":1760832000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["62431017,62320106003,U24A20251,62125109,62371288,62301299,62401357,62401366,62120106007"],"award-info":[{"award-number":["62431017,62320106003,U24A20251,62125109,62371288,62301299,62401357,62401366,62120106007"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2025,10,19]]},"DOI":"10.1109\/iccv51701.2025.01996","type":"proceedings-article","created":{"date-parts":[[2026,4,29]],"date-time":"2026-04-29T19:45:49Z","timestamp":1777491949000},"page":"21492-21504","source":"Crossref","is-referenced-by-count":0,"title":["METEOR: Multi-Encoder Collaborative Token Pruning for Efficient Vision Language Models"],"prefix":"10.1109","author":[{"given":"Yuchen","family":"Liu","sequence":"first","affiliation":[{"name":"Shanghai Jiao Tong University,China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yaoming","family":"Wang","sequence":"additional","affiliation":[{"name":"Meituan Inc.,China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Bowen","family":"Shi","sequence":"additional","affiliation":[{"name":"Shanghai Jiao Tong University,China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Xiaopeng","family":"Zhang","sequence":"additional","affiliation":[{"name":"Huawei Inc.,China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Wenrui","family":"Dai","sequence":"additional","affiliation":[{"name":"Shanghai Jiao Tong University,China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Chenglin","family":"Li","sequence":"additional","affiliation":[{"name":"Shanghai Jiao Tong University,China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Hongkai","family":"Xiong","sequence":"additional","affiliation":[{"name":"Shanghai Jiao Tong University,China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Qi","family":"Tian","sequence":"additional","affiliation":[{"name":"Huawei Inc.,China"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"263","reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v39i2.32171"},{"key":"ref2","article-title":"Qwen technical report","author":"Bai","year":"2023","journal-title":"arXiv preprint"},{"key":"ref3","article-title":"Qwen-VL: A frontier large vision-language model with versatile abilities","author":"Bai","year":"2023","journal-title":"arXiv preprint"},{"key":"ref4","first-page":"1877","article-title":"Language models are few-shot learners","author":"Brown","year":"2020","journal-title":"Advances in Neural Information Processing Systems 33"},{"key":"ref5","article-title":"Matryoshka multi-modal models","volume-title":"The Thirteenth International Conference on Learning Representations","author":"Cai","year":"2025"},{"key":"ref6","volume-title":"Open-LLaVA-NeXT: An opensource implementation of LLaVA-NeXT series for facilitating the large multi-modal model community","author":"Chen","year":"2024"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-73004-7_2"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1007\/s11432-024-4231-5"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.02283"},{"key":"ref10","author":"Chiang","year":"2023","journal-title":"Vicuna: An open-source chatbot impressing GPT-4 with 90%* ChatGPT quality"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.52202\/075280-2142"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/P19-1285"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.01554"},{"key":"ref14","article-title":"MouSi: Poly-visual-expert vision-language models","author":"Fan","year":"2024","journal-title":"arXiv preprint"},{"key":"ref15","article-title":"MME: A comprehensive evaluation benchmark for multi-modal large language models","author":"Fu","year":"2023","journal-title":"arXiv preprint"},{"key":"ref16","first-page":"390","article-title":"Llava-uhd: an 1 mm perceiving any aspect ratio and highresolution images","volume-title":"Proceedings of the 18th European Conference on Computer Vision (ECCV)","author":"Guo","year":"2024"},{"key":"ref17","article-title":"Rethinking token reduction in MLLMs: Towards a unified paradigm for training-free acceleration","author":"Han","year":"2024","journal-title":"arXiv preprint"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.24963\/ijcai.2024\/123"},{"key":"ref19","article-title":"ZipVL: Efficient large vision-language models with dynamic token sparsification and KV cache compression","author":"He","year":"2024","journal-title":"arXiv preprint"},{"key":"ref20","article-title":"iLLaVA: An image is worth fewer than 1\/3 input tokens in large multi-modal models","author":"Hu","year":"2024","journal-title":"arXiv preprint"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.52202\/079017-1588"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-72643-9_13"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2019.00686"},{"key":"ref24","article-title":"From CLIP to DINO: Visual encoders shout in multi-modal large language models","author":"Jiang","year":"2023","journal-title":"arXiv preprint"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2025.emnlp-main.1074"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52734.2025.02328"},{"key":"ref27","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-72640-8_7"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-46493-0_15"},{"key":"ref29","first-page":"273","article-title":"MoAI: Mixture of all intelligence for large language and vision models","volume-title":"Proceedings of the 18th European Conference on Computer Vision (ECCV)","author":"Lee","year":"2024"},{"key":"ref30","first-page":"18893","article-title":"Pix2Struct: Screenshot parsing as pretraining for visual language understanding","volume-title":"Proceedings of the 40th International Conference on Machine Learning (ICML)","author":"Lee","year":"2023"},{"key":"ref31","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.01263"},{"key":"ref32","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2025.findings-acl.1233"},{"key":"ref33","doi-asserted-by":"publisher","DOI":"10.1007\/s11263-025-02491-7"},{"key":"ref34","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2023.emnlp-main.20"},{"key":"ref35","doi-asserted-by":"publisher","DOI":"10.1109\/tpami.2025.3637265"},{"key":"ref36","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.02527"},{"key":"ref37","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR42600.2020.00160"},{"key":"ref38","first-page":"36","article-title":"SPHINX: A mixer of weights, visual embeddings and image scales for multi-modal large language models","volume-title":"Proceedings of the 18th European Conference on Computer Vision (ECCV)","author":"Lin","year":"2024"},{"key":"ref39","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v39i5.32567"},{"key":"ref40","doi-asserted-by":"publisher","DOI":"10.52202\/075280-1516"},{"key":"ref41","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.02484"},{"key":"ref42","author":"Liu","year":"2024","journal-title":"LLaVA- NeXT: Improved reasoning, OCR, and world knowledge"},{"key":"ref43","article-title":"Prismer: A vision-language model with multi-task experts","author":"Liu","year":"2024","journal-title":"Transactions on Machine Learning Resear"},{"key":"ref44","article-title":"Multi-stage vision token dropping: Towards efficient multi-modal large language model","author":"Liu","year":"2024","journal-title":"arXiv preprint"},{"key":"ref45","article-title":"On the hidden mystery of OCR in large multi-modal models","author":"Liu","year":"2023","journal-title":"arXiv preprint"},{"key":"ref46","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-72658-3_13"},{"key":"ref47","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52688.2022.01167"},{"key":"ref48","article-title":"DeepSeek-VL: towards real-world visionlanguage understanding","author":"Lu","year":"2024","journal-title":"arXiv preprint"},{"key":"ref49","doi-asserted-by":"publisher","DOI":"10.52202\/068431-0182"},{"key":"ref50","article-title":"Feast your eyes: Mixture-of-resolution adaptation for multi-modal large language mod-els","volume-title":"The Thirteenth International Conference on Learning Representations","author":"Luo","year":"2025"},{"key":"ref51","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2019.00331"},{"key":"ref52","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2022.findings-acl.177"},{"key":"ref53","doi-asserted-by":"publisher","DOI":"10.1109\/WACV48630.2021.00225"},{"key":"ref54","first-page":"304","article-title":"MM1: methods, analysis and insights from multi-modal LLM pretraining","volume-title":"Proceedings of the 18th European Conference on Computer Vision (ECCV)","author":"McKinzie","year":"2024"},{"key":"ref55","doi-asserted-by":"publisher","DOI":"10.52202\/079017-0739"},{"key":"ref56","first-page":"8748","article-title":"Learning transferable visual models from natural language supervision","volume-title":"Proceedings of the 38th International Conference on Machine Learning (ICML)","author":"Radford","year":"2021"},{"key":"ref57","article-title":"LLaVA- PruMerge: Adaptive token reduction for efficient large multi-modal models","author":"Shang","year":"2024","journal-title":"arXiv preprint"},{"key":"ref58","article-title":"Eagle: Exploring the design space for multi-modal LLMs with mixture of encoders","volume-title":"The Thirteenth International Conference on Learning Representations","author":"Shi","year":"2025"},{"key":"ref59","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2019.00851"},{"key":"ref60","article-title":"EVA-CLIP: Improved training techniques for CLIP at scale","author":"Sun","year":"2023","journal-title":"arXiv preprint"},{"key":"ref61","doi-asserted-by":"publisher","DOI":"10.52202\/079017-2771"},{"key":"ref62","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.00914"},{"key":"ref63","article-title":"LLaMA: Open and efficient foundation language models","author":"Touvron","year":"2023","journal-title":"arXiv preprint"},{"key":"ref64","article-title":"[CLS] token tells everything needed for training-free efficient MLLMs","author":"Wang","year":"2024","journal-title":"arXiv preprint"},{"key":"ref65","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51701.2025.02192"},{"key":"ref66","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-73235-5_23"},{"key":"ref67","article-title":"PyramidDrop: Accelerating your large vision-language models via pyramid visual redundancy reduction","author":"Xing","year":"2024","journal-title":"arXiv preprint"},{"key":"ref68","article-title":"Mitigating hallucination in large vision-language models via modular attribution and intervention","volume-title":"The Thirteenth International Conference on Learning Representations","author":"Yang","year":"2025"},{"key":"ref69","article-title":"DeCo: Decoupling token compression from semantic abstraction in multi-modal large language models","author":"Yao","year":"2024","journal-title":"arXiv preprint"},{"key":"ref70","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.01239"},{"key":"ref71","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v39i21.34366"},{"key":"ref72","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52734.2025.02325"},{"key":"ref73","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52734.2025.02777"},{"key":"ref74","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52734.2025.00876"},{"key":"ref75","article-title":"[CLS] attention is all you need for training-free visual token pruning: Make VLM inference faster","author":"Zhang","year":"2024","journal-title":"arXiv preprint"},{"key":"ref76","article-title":"LLaVA- Mini: Efficient image and video large multi-modal models with one vision token","volume-title":"The Thirteenth International Conference on Learning Representations","author":"Zhang","year":"2025"},{"key":"ref77","article-title":"Seeing clearly by layer two: Enhancing attention heads to alleviate hallucination in LVLMs","author":"Zhang","year":"2024","journal-title":"arXiv preprint"},{"key":"ref78","article-title":"SparseVLM: Visual token sparsification for efficient visionlanguage model inference","volume-title":"Proceedings of the 42nd International Conference on Machine Learning","author":"Zhang","year":"2025"},{"key":"ref79","article-title":"Treat visual tokens as text? but your MLLM only needs fewer efforts to see","author":"Zhang","year":"2024","journal-title":"arXiv preprint"},{"key":"ref80","article-title":"Accelerating multimodel large language models by searching optimal vision token reduction","author":"Zhao","year":"2024","journal-title":"arXiv preprint"},{"key":"ref81","article-title":"AIM: Adaptive inference of multi-modal LLMs via token merging and pruning","author":"Zhong","year":"2024","journal-title":"arXiv preprint"},{"key":"ref82","article-title":"MiniGPT-4: Enhancing vision-language understanding with advanced large language models","author":"Zhu","year":"2023","journal-title":"arXiv preprint"},{"key":"ref83","article-title":"FocusLLaVA: A coarse-to-fine approach for effi-cient and effective visual token compression","author":"Zhu","year":"2024","journal-title":"arXiv preprint"},{"key":"ref84","doi-asserted-by":"publisher","DOI":"10.52202\/079017-3282"}],"event":{"name":"2025 IEEE\/CVF International Conference on Computer Vision (ICCV)","location":"Honolulu, HI, USA","start":{"date-parts":[[2025,10,19]]},"end":{"date-parts":[[2025,10,25]]}},"container-title":["2025 IEEE\/CVF International Conference on Computer Vision (ICCV)"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/11443115\/11443287\/11446273.pdf?arnumber=11446273","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,4,30]],"date-time":"2026-04-30T06:25:29Z","timestamp":1777530329000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11446273\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,10,19]]},"references-count":84,"URL":"https:\/\/doi.org\/10.1109\/iccv51701.2025.01996","relation":{},"subject":[],"published":{"date-parts":[[2025,10,19]]}}}