{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,4]],"date-time":"2026-07-04T17:11:13Z","timestamp":1783185073634,"version":"3.54.6"},"reference-count":103,"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"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2025,10,19]]},"DOI":"10.1109\/iccv51701.2025.02021","type":"proceedings-article","created":{"date-parts":[[2026,4,29]],"date-time":"2026-04-29T19:45:49Z","timestamp":1777491949000},"page":"21766-21777","source":"Crossref","is-referenced-by-count":4,"title":["Describe Anything: Detailed Localized Image and Video Captioning"],"prefix":"10.1109","author":[{"given":"Long","family":"Lian","sequence":"first","affiliation":[{"name":"NVIDIA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yifan","family":"Ding","sequence":"additional","affiliation":[{"name":"NVIDIA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yunhao","family":"Ge","sequence":"additional","affiliation":[{"name":"NVIDIA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Sifei","family":"Liu","sequence":"additional","affiliation":[{"name":"NVIDIA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hanzi","family":"Mao","sequence":"additional","affiliation":[{"name":"NVIDIA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Boyi","family":"Li","sequence":"additional","affiliation":[{"name":"NVIDIA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Marco","family":"Pavone","sequence":"additional","affiliation":[{"name":"NVIDIA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ming-Yu","family":"Liu","sequence":"additional","affiliation":[{"name":"NVIDIA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Trevor","family":"Darrell","sequence":"additional","affiliation":[{"name":"NVIDIA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Adam","family":"Yala","sequence":"additional","affiliation":[{"name":"UC Berkeley"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yin","family":"Cui","sequence":"additional","affiliation":[{"name":"NVIDIA"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.52202\/068431-1723"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-46454-1_24"},{"key":"ref3","author":"Awadalla","year":"2023","journal-title":"Openflamingo: An open-source framework for training large autoregressive vision-language models"},{"key":"ref4","author":"Bai","year":"2023","journal-title":"Qwen technical report"},{"key":"ref5","article-title":"Meteor: An automatic metric for mt evaluation with improved correlation with human judgments","volume-title":"ACL Workshops","author":"Banerjee","year":"2005"},{"key":"ref6","author":"Bavishi","year":"2023","journal-title":"Fuyu-8b: A multimodal architecture for ai agents"},{"key":"ref7","author":"Berthelot","year":"2019","journal-title":"Remixmatch: Semi-supervised learning with distribution alignment and augmentation anchoring"},{"key":"ref8","article-title":"Mixmatch: A holistic approach to semi-supervised learning","author":"Berthelot","year":"2019","journal-title":"NeurIPS"},{"key":"ref9","author":"Beyer","year":"2024","journal-title":"Paligemma: A versatile 3b vlm for transfer"},{"key":"ref10","author":"Bolya","year":"2022","journal-title":"Token merging: Your vit but faster"},{"key":"ref11","article-title":"Coco-stuff: Thing and stuff classes in context","author":"Caesar","year":"2016","journal-title":"arxiv"},{"key":"ref12","article-title":"Cascade r-cnn: High quality object detection and instance segmentation","author":"Cai","year":"2019","journal-title":"TPAMI"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2023.emnlp-main.841"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1109\/cvprw67362.2025.00056"},{"key":"ref15","author":"Chen","year":"2023","journal-title":"Shikra: Unleashing multimodal llm\u2019s referential dialogue magic"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-72643-9_22"},{"key":"ref17","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.01265"},{"key":"ref18","author":"Chen","year":"2015","journal-title":"Microsoft coco captions: Data collection and evaluation server"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.01368"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1109\/cvpr52733.2024.02283"},{"key":"ref21","author":"Chen","year":"2024","journal-title":"Expanding performance boundaries of open-source multimodal models with model, data, and test-time scaling"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.52202\/075280-2142"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.01850"},{"key":"ref24","author":"Dosovitskiy","year":"2020","journal-title":"An image is worth 16 x 16 words: Transformers for image recognition at scale"},{"key":"ref25","author":"Dubey","year":"2024","journal-title":"The llama 3 herd of models"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.1016\/j.imavis.2024.105171"},{"key":"ref27","article-title":"Dataseg: Taming a universal multi-dataset multi-task segmentation model","author":"Gu","year":"2024","journal-title":"NeurIPS"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.01309"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2019.00550"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.01273"},{"key":"ref31","doi-asserted-by":"publisher","DOI":"10.3115\/v1\/D14-1086"},{"key":"ref32","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.00371"},{"key":"ref33","article-title":"Openimages: A public dataset for large-scale multi-label and multiclass image classification","volume-title":"Dataset","author":"Krasin","year":"2017"},{"key":"ref34","doi-asserted-by":"publisher","DOI":"10.1007\/s11263-016-0981-7"},{"key":"ref35","doi-asserted-by":"publisher","DOI":"10.1007\/s11263-020-01316-z"},{"key":"ref36","article-title":"Obelics: An open web-scale filtered dataset of interleaved image-text documents","author":"Lauren\u00e7on","year":"2024","journal-title":"NeurIPS"},{"key":"ref37","first-page":"896","article-title":"Pseudo-label: The simple and efficient semi-supervised learning method for deep neural networks","volume-title":"In Workshop on challenges in representation learning, ICML","author":"Lee"},{"key":"ref38","article-title":"Blip: Bootstrapping language-image pre-training for unified vision-language understanding and generation","author":"Li","year":"2022","journal-title":"In ICML"},{"key":"ref39","article-title":"Blip-2: Bootstrapping language-image pre-training with frozen image encoders and large language models","author":"Li","year":"2023","journal-title":"ICML"},{"key":"ref40","first-page":"10955","article-title":"Grounded language-image pre-training. 2022 ieee","author":"Harold Li","year":"2021","journal-title":"In CVPR"},{"key":"ref41","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-20077-9_17"},{"key":"ref42","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52729.2023.02156"},{"key":"ref43","first-page":"74","article-title":"Rouge: A package for automatic evaluation of summaries","author":"Lin","year":"2004","journal-title":"Text summarization branches out"},{"key":"ref44","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.02520"},{"key":"ref45","article-title":"Draw-and-understand: Leveraging visual prompts to enable MLLMs to comprehend what you want","author":"Lin","year":"2025","journal-title":"In ICLR"},{"key":"ref46","author":"Liu","year":"2023","journal-title":"Improved baselines with visual instruction tuning"},{"key":"ref47","author":"Liu","year":"2024","journal-title":"Llava-next: Improved reasoning, ocr, and world knowledge"},{"key":"ref48","article-title":"Visual instruction tuning","author":"Liu","year":"2024","journal-title":"NeurIPS"},{"key":"ref49","first-page":"38","article-title":"Grounding dino: Marrying dino with grounded pre-training for open-set object detection","volume-title":"European Conference on Computer Vision","author":"Liu"},{"key":"ref50","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.9"},{"key":"ref51","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-20080-9_42"},{"key":"ref52","article-title":"Scaling open-vocabulary object detection","author":"Minderer","year":"2024","journal-title":"NeurIPS"},{"key":"ref53","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2017.534"},{"key":"ref54","article-title":"Gpt-4o system card","year":"2024","journal-title":"OpenAI"},{"key":"ref55","article-title":"Learning to reason with 11 ms","year":"2024","journal-title":"OpenAI"},{"key":"ref56","first-page":"311","article-title":"Bleu: a method for automatic evaluation of machine translation","author":"Papineni","year":"2002","journal-title":"A C L"},{"key":"ref57","author":"Peng","year":"2023","journal-title":"Kosmos-2: Grounding multimodal large language models to the world"},{"key":"ref58","first-page":"2641","article-title":"Flickr30k entities: Collecting region-to-phrase correspondences for richer image-to-sentence models","author":"Bryan","year":"2015","journal-title":"In ICCV"},{"key":"ref59","first-page":"114321","article-title":"Artemis: Towards referential understanding in complex videos","volume":"37","author":"Qiu","year":"2024","journal-title":"NeurIPS"},{"key":"ref60","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52729.2023.00690"},{"key":"ref61","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.01236"},{"key":"ref62","author":"Ravi","year":"2024","journal-title":"Sam 2: Segment anything in images and videos"},{"key":"ref63","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/D19-1410"},{"key":"ref64","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV48922.2021.01072"},{"key":"ref65","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52688.2022.01168"},{"key":"ref66","doi-asserted-by":"publisher","DOI":"10.1109\/CVPRW59228.2023.00486"},{"key":"ref67","article-title":"Objects 365: A large-scale, high-quality dataset for object detection","author":"Shao","year":"2019","journal-title":"In ICCV"},{"key":"ref68","author":"Kumar Singh","year":"2024","journal-title":"Tosa: Token selective attention for efficient vision transformers"},{"key":"ref69","article-title":"Fixmatch: Simplifying semi-supervised learning with consistency and confidence","author":"Sohn","year":"2020","journal-title":"NeurIPS"},{"key":"ref70","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.01237"},{"key":"ref71","doi-asserted-by":"publisher","DOI":"10.1109\/TCSVT.2021.3085907"},{"key":"ref72","article-title":"Claude 3.5 sonnet","year":"2024","journal-title":"Claude Team."},{"key":"ref73","article-title":"Claude 3.7 sonnet and claude code","year":"2025","journal-title":"Claude Team."},{"key":"ref74","article-title":"Gemini 2.5: Our most intelligent ai model","year":"2025","journal-title":"Gemini Team."},{"key":"ref75","author":"Team","year":"2023","journal-title":"Gemini: a family of highly capable multimodal models"},{"key":"ref76","author":"Team","year":"2024","journal-title":"Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context"},{"key":"ref77","article-title":"Qwen2.5-vl","year":"2025","journal-title":"Qwen Team."},{"key":"ref78","article-title":"Attention is all you need","author":"Vaswani","year":"2017","journal-title":"NeurIPS"},{"key":"ref79","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2015.7299087"},{"key":"ref80","first-page":"166","article-title":"Elysium: Exploring object-level perception in videos via mllm","volume-title":"In ECCV","author":"Wang","year":"2024"},{"key":"ref81","author":"Wang","year":"2024","journal-title":"Qwen2-vl: Enhancing visionlanguage model\u2019s perception of the world at any resolution"},{"key":"ref82","author":"Wang","year":"2023","journal-title":"Caption anything: Interactive image description with diverse multimodal controls"},{"key":"ref83","author":"Wang","year":"2023","journal-title":"The all-seeing project: Towards panoptic visual recognition and understanding of the open world"},{"key":"ref84","author":"Wang","year":"2024","journal-title":"Enhancing the reasoning ability of multimodal large language models via mixed preference optimization"},{"key":"ref85","author":"Wu","year":"2022","journal-title":"Grit: A generative region-to-text transformer for object understanding"},{"key":"ref86","doi-asserted-by":"publisher","DOI":"10.1109\/cvpr52733.2024.01243"},{"key":"ref87","author":"Wu","year":"2024","journal-title":"Vila-u: a unified foundation model integrating visual understanding and generation"},{"key":"ref88","article-title":"Unsupervised data augmentation for consistency training","author":"Xie","year":"2020","journal-title":"NeurIPS"},{"key":"ref89","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR42600.2020.01070"},{"key":"ref90","author":"Yang","year":"2021","journal-title":"Focal selfattention for local-global interactions in vision transformers"},{"key":"ref91","author":"Yang","year":"2023","journal-title":"Set-of-mark prompting unleashes extraordinary visual grounding in gpt-4v"},{"key":"ref92","author":"Yang","year":"2023","journal-title":"The dawn of 1 mms: Preliminary explorations with gpt-4v (ision)"},{"key":"ref93","doi-asserted-by":"publisher","DOI":"10.1036\/1097-8542.253500"},{"key":"ref94","first-page":"425","article-title":"Merlin: Empowering multimodal 11 ms with foresight minds","volume-title":"In ECCV","author":"Yu","year":"2024"},{"key":"ref95","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.02664"},{"key":"ref96","doi-asserted-by":"publisher","DOI":"10.1109\/cvpr52734.2025.01767"},{"key":"ref97","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.01100"},{"key":"ref98","article-title":"Ferret-v2: An improved baseline for referring and grounding with large language models","author":"Zhang","year":"2024","journal-title":"COLM"},{"key":"ref99","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-91813-1_4"},{"key":"ref100","author":"Zhang","year":"2024","journal-title":"Omg-llava: Bridging image-level, object-level, pixel-level reasoning and understanding"},{"key":"ref101","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-72920-1_2"},{"key":"ref102","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52688.2022.01629"},{"key":"ref103","author":"Zhu","year":"2023","journal-title":"Minigpt-4: Enhancing visionlanguage understanding with advanced large language models"}],"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\/11443506.pdf?arnumber=11443506","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,4,30]],"date-time":"2026-04-30T06:17:05Z","timestamp":1777529825000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11443506\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,10,19]]},"references-count":103,"URL":"https:\/\/doi.org\/10.1109\/iccv51701.2025.02021","relation":{},"subject":[],"published":{"date-parts":[[2025,10,19]]}}}