{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,12]],"date-time":"2026-05-12T16:49:57Z","timestamp":1778604597176,"version":"3.51.4"},"reference-count":115,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","issue":"11","license":[{"start":{"date-parts":[[2025,11,1]],"date-time":"2025-11-01T00:00:00Z","timestamp":1761955200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2025,11,1]],"date-time":"2025-11-01T00:00:00Z","timestamp":1761955200000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2025,11,1]],"date-time":"2025-11-01T00:00:00Z","timestamp":1761955200000},"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":["IEEE Trans. Pattern Anal. Mach. Intell."],"published-print":{"date-parts":[[2025,11]]},"DOI":"10.1109\/tpami.2025.3596776","type":"journal-article","created":{"date-parts":[[2025,8,7]],"date-time":"2025-08-07T17:44:12Z","timestamp":1754588652000},"page":"10560-10576","source":"Crossref","is-referenced-by-count":5,"title":["EMOv2: Pushing 5M Vision Model Frontier"],"prefix":"10.1109","volume":"47","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-8891-6766","authenticated-orcid":false,"given":"Jiangning","family":"Zhang","sequence":"first","affiliation":[{"name":"Institute of Cyber-Systems and Control, Zhejiang University, Hangzhou, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0009-0008-1247-5931","authenticated-orcid":false,"given":"Teng","family":"Hu","sequence":"additional","affiliation":[{"name":"Shanghai Jiao Tong University, Shanghai, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0009-0003-2771-5196","authenticated-orcid":false,"given":"Haoyang","family":"He","sequence":"additional","affiliation":[{"name":"Institute of Cyber-Systems and Control, Zhejiang University, Hangzhou, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0009-0005-2627-4397","authenticated-orcid":false,"given":"Zhucun","family":"Xue","sequence":"additional","affiliation":[{"name":"Institute of Cyber-Systems and Control, Zhejiang University, Hangzhou, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yabiao","family":"Wang","sequence":"additional","affiliation":[{"name":"Youtu Lab, Tencent, Shenzhen, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-4216-8090","authenticated-orcid":false,"given":"Chengjie","family":"Wang","sequence":"additional","affiliation":[{"name":"Youtu Lab, Tencent, Shenzhen, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-4822-8939","authenticated-orcid":false,"given":"Yong","family":"Liu","sequence":"additional","affiliation":[{"name":"Institute of Cyber-Systems and Control, Zhejiang University, Hangzhou, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-0550-8247","authenticated-orcid":false,"given":"Xiangtai","family":"Li","sequence":"additional","affiliation":[{"name":"Nanyang Technological University, Singapore"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7225-5449","authenticated-orcid":false,"given":"Dacheng","family":"Tao","sequence":"additional","affiliation":[{"name":"Nanyang Technological University, Singapore"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"263","reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.01549"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-25082-8_1"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v39i19.34255"},{"key":"ref4","article-title":"EdgeSAM: Prompt-in-the-loop distillation for on-device deployment of SAM","author":"Zhou","year":"2023"},{"key":"ref5","article-title":"RMP-SAM: Towards real-time multi-purpose segment anything","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Xu"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-58452-8_45"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.00148"},{"key":"ref8","article-title":"MobileNets: Efficient convolutional neural networks for mobile vision applications","author":"Howard","year":"2017"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2018.00474"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2019.00140"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR42600.2020.00165"},{"key":"ref12","first-page":"6105","article-title":"EfficientNet: Rethinking model scaling for convolutional neural networks","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Tan"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.00134"},{"key":"ref14","article-title":"Separable self-attention for mobile vision transformers","author":"Mehta","year":"2023","journal-title":"Trans. Mach. Learn. Res."},{"key":"ref15","article-title":"The need for speed in AI","author":"Nielsen","year":"2023"},{"key":"ref16","volume-title":"Usability Engineering","author":"Nielsen","year":"1994"},{"key":"ref17","article-title":"MobileViT: Light-weight, general-purpose, and mobile-friendly vision transformer","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Mehta"},{"key":"ref18","article-title":"An image is worth 16x16 words: Transformers for image recognition at scale","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Dosovitskiy"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV48922.2021.00061"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1007\/s41095-022-0274-8"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV48922.2021.00986"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52688.2022.01170"},{"key":"ref23","first-page":"26674","article-title":"Analogous to evolutionary algorithm: Designing a unified sequence model","volume-title":"Proc. Int. Conf. Neural Inf. Process. Syst.","author":"Zhang"},{"key":"ref24","doi-asserted-by":"publisher","DOI":"10.1007\/s11263-024-02034-6"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2024.3434373"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR46437.2021.01214"},{"key":"ref27","article-title":"Reformer: The efficient transformer","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Kitaev"},{"key":"ref28","article-title":"Rethinking attention with performers","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Choromanski"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV48922.2021.00009"},{"key":"ref30","article-title":"Next-ViT: Next generation vision transformer for efficient deployment in realistic industrial scenarios","author":"Li","year":"2022"},{"key":"ref31","article-title":"DeLighT: Deep and light-weight transformer","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Mehta"},{"key":"ref32","article-title":"MobileViTv3: Mobile-friendly vision transformer with simple and effective fusion of local, global and input features","author":"Wadekar","year":"2022"},{"key":"ref33","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52688.2022.00520"},{"key":"ref34","article-title":"EfficientFormer: Vision transformers at MobileNet speed","volume-title":"Proc. Int. Conf. Neural Inf. Process. Syst.","author":"Li"},{"key":"ref35","doi-asserted-by":"publisher","DOI":"10.48550\/ARXIV.1706.03762"},{"key":"ref36","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2017.324"},{"key":"ref37","article-title":"SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and $< $<0.5 MB model size","author":"Iandola","year":"2016"},{"key":"ref38","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.308"},{"key":"ref39","doi-asserted-by":"publisher","DOI":"10.1007\/s11263-023-01875-x"},{"key":"ref40","doi-asserted-by":"publisher","DOI":"10.1109\/cvpr52733.2024.01506"},{"key":"ref41","article-title":"GhostNetV3: Exploring the training strategies for compact models","author":"Liu","year":"2024"},{"key":"ref42","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-73661-2_5"},{"key":"ref43","first-page":"10347","article-title":"Training data-efficient image transformers & distillation through attention","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Touvron"},{"key":"ref44","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.90"},{"key":"ref45","doi-asserted-by":"publisher","DOI":"10.1016\/j.inffus.2024.102417"},{"key":"ref46","article-title":"Recent advances in vision transformer: A survey and outlook of recent work","author":"Islam","year":"2022"},{"key":"ref47","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV48922.2021.00062"},{"key":"ref48","article-title":"Conditional positional encodings for vision transformers","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Chu"},{"key":"ref49","article-title":"UniFormer: Unified transformer for efficient spatial-temporal representation learning","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Li"},{"key":"ref50","article-title":"MogaNet: Multi-order gated aggregation network","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Li"},{"key":"ref51","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.00550"},{"key":"ref52","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52688.2022.01055"},{"key":"ref53","article-title":"LightViT: Towards light-weight convolution-free vision transformers","author":"Huang","year":"2022"},{"key":"ref54","article-title":"ResT: An efficient transformer for visual recognition","volume-title":"Proc. Int. Conf. Neural Inf. Process. Syst.","author":"Zhang"},{"key":"ref55","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-20083-0_18"},{"key":"ref56","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2019.2938758"},{"key":"ref57","first-page":"20014","article-title":"XCiT: Cross-covariance image transformers","volume-title":"Proc. Int. Conf. Neural Inf. Process. Syst.","author":"Ali"},{"key":"ref58","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v39i5.32550"},{"key":"ref59","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v39i3.32353"},{"key":"ref60","article-title":"Vision-RWKV: Efficient and scalable visual perception with RWKV-like architectures","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Duan"},{"key":"ref61","article-title":"Mamba or RWKV: Exploring high-quality and high-efficiency segment anything model","author":"Yuan","year":"2024"},{"key":"ref62","article-title":"Mamba: Linear-time sequence modeling with selective state spaces","volume-title":"Proc. 1st Conf. Lang. Model.","author":"Gu"},{"key":"ref63","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2023.findings-emnlp.936"},{"key":"ref64","article-title":"Vision Mamba: Efficient visual representation learning with bidirectional state space model","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Zhu"},{"key":"ref65","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v39i6.32690"},{"key":"ref66","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52734.2025.00424"},{"key":"ref67","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.00387"},{"key":"ref68","first-page":"30008","article-title":"Focal attention for long-range interactions in vision transformers","volume-title":"Proc. Int. Conf. Neural Inf. Process. Syst.","author":"Yang"},{"key":"ref69","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52688.2022.01181"},{"key":"ref70","article-title":"Inception transformer","volume-title":"Proc. Int. Conf. Neural Inf. Process. Syst.","author":"Si"},{"key":"ref71","first-page":"9204","article-title":"Pay attention to MLPs","volume-title":"Proc. Int. Conf. Neural Inf. Process. Syst.","author":"Liu"},{"key":"ref72","first-page":"24261","article-title":"MLP-Mixer: An all-MLP architecture for vision","volume-title":"Proc. Int. Conf. Neural Inf. Process. Syst.","author":"Tolstikhin"},{"key":"ref73","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2022.3206148"},{"key":"ref74","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-01264-9_8"},{"key":"ref75","article-title":"MOAT: Alternating mobile convolution and attention brings strong vision models","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Yang"},{"key":"ref76","doi-asserted-by":"publisher","DOI":"10.5555\/3045118.3045167"},{"key":"ref77","article-title":"Gaussian error linear units (GELUs)","author":"Hendrycks","year":"2016"},{"key":"ref78","article-title":"Layer normalization","author":"Ba","year":"2016"},{"key":"ref79","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2009.5206848"},{"key":"ref80","article-title":"Decoupled weight decay regularization","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Loshchilov"},{"key":"ref81","article-title":"SGDR: Stochastic gradient descent with warm restarts","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Loshchilov"},{"key":"ref82","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-46493-0_39"},{"key":"ref83","doi-asserted-by":"publisher","DOI":"10.1109\/CVPRW50498.2020.00359"},{"key":"ref84","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV48922.2021.00010"},{"key":"ref85","first-page":"1929","article-title":"Dropout: A simple way to prevent neural networks from overfitting","volume":"15","author":"Srivastava","year":"2014","journal-title":"J. Mach. Learn. Res."},{"key":"ref86","article-title":"mixup: Beyond empirical risk minimization","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Zhang"},{"key":"ref87","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2019.00612"},{"key":"ref88","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v34i07.7000"},{"key":"ref89","doi-asserted-by":"publisher","DOI":"10.48550\/ARXIV.2104.10858"},{"key":"ref90","article-title":"Pytorch image models","author":"Wightman","year":"2019"},{"key":"ref91","doi-asserted-by":"publisher","DOI":"10.1109\/WACV48630.2021.00144"},{"key":"ref92","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52729.2023.01157"},{"key":"ref93","article-title":"MoCoViT: Mobile convolutional vision transformer","author":"Ma","year":"2022"},{"key":"ref94","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52729.2023.01386"},{"key":"ref95","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52688.2022.00714"},{"key":"ref96","article-title":"Multi-scale VMamba: Hierarchy in hierarchy visual state space model","volume-title":"Proc. Int. Conf. Neural Inf. Process. Syst.","author":"Shi"},{"key":"ref97","doi-asserted-by":"publisher","DOI":"10.1109\/cvpr52734.2025.00423"},{"key":"ref98","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-10602-1_48"},{"key":"ref99","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2017.322"},{"key":"ref100","article-title":"MMDetection: Open MMLab detection toolbox and benchmark","author":"Chen","year":"2019"},{"key":"ref101","article-title":"Rethinking atrous convolution for semantic image segmentation","author":"Chen","year":"2017"},{"key":"ref102","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2019.00656"},{"key":"ref103","first-page":"12077","article-title":"SegFormer: Simple and efficient design for semantic segmentation with transformers","volume-title":"Proc. Int. Conf. Neural Inf. Process. Syst.","author":"Xie"},{"key":"ref104","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2017.660"},{"key":"ref105","doi-asserted-by":"publisher","DOI":"10.1007\/s11263-018-1140-0"},{"key":"ref106","article-title":"MMSegmentation: OpenMMLab semantic segmentation toolbox and benchmark","author":"Contributors","year":"2020"},{"key":"ref107","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-24574-4_28"},{"key":"ref108","doi-asserted-by":"publisher","DOI":"10.1155\/2013\/154860"},{"key":"ref109","article-title":"The Kinetics human action video dataset","author":"Kay","year":"2017"},{"key":"ref110","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-72980-5_2"},{"key":"ref111","article-title":"Optimize your core ML usage","year":"2022"},{"key":"ref112","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2019.00953"},{"key":"ref113","article-title":"ResNet strikes back: An improved training procedure in TIMM","volume-title":"Proc. Int. Conf. Neural Inf. Process. Syst. Workshops","author":"Wightman"},{"key":"ref114","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52688.2022.01167"},{"key":"ref115","doi-asserted-by":"publisher","DOI":"10.1007\/s11263-022-01739-w"}],"container-title":["IEEE Transactions on Pattern Analysis and Machine Intelligence"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/34\/11192800\/11119331.pdf?arnumber=11119331","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,10,6]],"date-time":"2025-10-06T17:36:34Z","timestamp":1759772194000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11119331\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,11]]},"references-count":115,"journal-issue":{"issue":"11"},"URL":"https:\/\/doi.org\/10.1109\/tpami.2025.3596776","relation":{},"ISSN":["0162-8828","2160-9292","1939-3539"],"issn-type":[{"value":"0162-8828","type":"print"},{"value":"2160-9292","type":"electronic"},{"value":"1939-3539","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,11]]}}}