{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,7,30]],"date-time":"2025-07-30T15:08:01Z","timestamp":1753888081867,"version":"3.41.2"},"reference-count":35,"publisher":"Wiley","issue":"1","license":[{"start":{"date-parts":[[2021,2,8]],"date-time":"2021-02-08T00:00:00Z","timestamp":1612742400000},"content-version":"vor","delay-in-days":38,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100008778","name":"University of Science and Technology Beijing","doi-asserted-by":"publisher","award":["FRF-DF-19-002"],"award-info":[{"award-number":["FRF-DF-19-002"]}],"id":[{"id":"10.13039\/501100008778","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["Complexity"],"published-print":{"date-parts":[[2021,1]]},"abstract":"<jats:p>A major challenge for semantic video segmentation is how to exploit the spatiotemporal information and produce consistent results for a video sequence. Many previous works utilize the precomputed optical flow to warp the feature maps across adjacent frames. However, the imprecise optical flow and the warping operation without any learnable parameters may not achieve accurate feature warping and only bring a slight improvement. In this paper, we propose a novel framework named Dynamic Warping Network (DWNet) to adaptively warp the interframe features for improving the accuracy of warping\u2010based models. Firstly, we design a flow refinement module (FRM) to optimize the precomputed optical flow. Then, we propose a flow\u2010guided convolution (FG\u2010Conv) to achieve the adaptive feature warping based on the refined optical flow. Furthermore, we introduce the temporal consistency loss including the feature consistency loss and prediction consistency loss to explicitly supervise the warped features instead of simple feature propagation and fusion, which guarantees the temporal consistency of video segmentation. Note that our DWNet adopts extra constraints to improve the temporal consistency in the training phase, while no additional calculation and postprocessing are required during inference. Extensive experiments show that our DWNet can achieve consistent improvement over various strong baselines and achieves state\u2010of\u2010the\u2010art accuracy on the Cityscapes and CamVid benchmark datasets.<\/jats:p>","DOI":"10.1155\/2021\/6680509","type":"journal-article","created":{"date-parts":[[2021,2,9]],"date-time":"2021-02-09T02:37:49Z","timestamp":1612838269000},"update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":7,"title":["Dynamic Warping Network for Semantic Video Segmentation"],"prefix":"10.1155","volume":"2021","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-2288-7901","authenticated-orcid":false,"given":"Jiangyun","family":"Li","sequence":"first","affiliation":[]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5923-3683","authenticated-orcid":false,"given":"Yikai","family":"Zhao","sequence":"additional","affiliation":[]},{"given":"Xingjian","family":"He","sequence":"additional","affiliation":[]},{"ORCID":"https:\/\/orcid.org\/0000-0002-2142-5580","authenticated-orcid":false,"given":"Xinxin","family":"Zhu","sequence":"additional","affiliation":[]},{"ORCID":"https:\/\/orcid.org\/0000-0003-0903-9131","authenticated-orcid":false,"given":"Jing","family":"Liu","sequence":"additional","affiliation":[]}],"member":"311","published-online":{"date-parts":[[2021,2,8]]},"reference":[{"key":"e_1_2_9_1_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2015.7298965"},{"key":"e_1_2_9_2_2","unstructured":"ChenL. C. PapandreouG. SchroffF. andAdamH. Rethinking atrous convolution for semantic image segmentation 2017."},{"key":"e_1_2_9_3_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-01234-2_49"},{"key":"e_1_2_9_4_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2018.00747"},{"key":"e_1_2_9_5_2","first-page":"3146","article-title":"Dual attention network for scene segmentation","volume":"22","author":"Fu J.","year":"2019","journal-title":"Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition"},{"key":"e_1_2_9_6_2","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2017.477"},{"key":"e_1_2_9_7_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2017.441"},{"volume-title":"Accel: A Corrective Fusion Network for Efficient Semantic Segmentation on Video","year":"2019","author":"Jain S.","key":"e_1_2_9_8_2"},{"volume-title":"How to Train Your Dragon: Tamed Warping Network for Semantic Video Segmentation","year":"2020","author":"Feng J.","key":"e_1_2_9_9_2"},{"volume-title":"Flownet: Learning Optical Flow with Convolutional Networks","year":"2015","author":"Dosovitskiy A.","key":"e_1_2_9_10_2"},{"key":"e_1_2_9_11_2","doi-asserted-by":"crossref","unstructured":"IlgE. MayerN. SaikiaT. KeuperM. DosovitskiyA. andBroxT. Flownet 2.0: evolution of optical flow estimation with deep networks 2017 CVPR London UK.","DOI":"10.1109\/CVPR.2017.179"},{"volume-title":"Pwc-net: Cnns for Optical Flow Using Pyramid, Warping, and Cost Volume","year":"2018","author":"Sun D.","key":"e_1_2_9_12_2"},{"volume-title":"Every Frame Counts: Joint Learning of Video Segmentation and Optical Flow","year":"2020","author":"Ding M.","key":"e_1_2_9_13_2"},{"key":"e_1_2_9_14_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2018.00929"},{"key":"e_1_2_9_15_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2018.00713"},{"key":"e_1_2_9_16_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2018.00686"},{"key":"e_1_2_9_17_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-58607-2_21"},{"key":"e_1_2_9_18_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2018.00628"},{"key":"e_1_2_9_19_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR42600.2020.00884"},{"key":"e_1_2_9_20_2","unstructured":"FayyazM. SaffarM. H. SabokrouM. FathyM. KletteR. andHuangF. STFCN: spatio-temporal FCN for semantic video segmentation 2016."},{"key":"e_1_2_9_21_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPRW.2016.57"},{"volume-title":"Bringing Background into the Foreground: Making All Classes Equal in Weakly-Supervised Video Semantic Segmentation","year":"2017","author":"Saleh F. S.","key":"e_1_2_9_22_2"},{"key":"e_1_2_9_23_2","unstructured":"ZhuangJ. WangZ. andWangB. Video semantic segmentation with distortion-aware feature correction 2020."},{"volume-title":"Dynamic Filter Networks","year":"2016","author":"Bert D. B.","key":"e_1_2_9_24_2"},{"key":"e_1_2_9_25_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2019.01142"},{"key":"e_1_2_9_26_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-58595-2_46"},{"key":"e_1_2_9_27_2","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2017.89"},{"volume-title":"Deformable ConvNets V2: More Deformable, Better Results","year":"2019","author":"Zhu X.","key":"e_1_2_9_28_2"},{"key":"e_1_2_9_29_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.350"},{"key":"e_1_2_9_30_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-540-88682-2_5"},{"key":"e_1_2_9_31_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.90"},{"volume-title":"MobileNetV2: Inverted Residuals and Linear Bottlenecks","year":"2018","author":"Sandler M.","key":"e_1_2_9_32_2"},{"key":"e_1_2_9_33_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2017.660"},{"key":"e_1_2_9_34_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-49409-8_69"},{"key":"e_1_2_9_35_2","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2017.595"}],"container-title":["Complexity"],"original-title":[],"language":"en","link":[{"URL":"http:\/\/downloads.hindawi.com\/journals\/complexity\/2021\/6680509.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"http:\/\/downloads.hindawi.com\/journals\/complexity\/2021\/6680509.xml","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1155\/2021\/6680509","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,8,9]],"date-time":"2024-08-09T22:31:14Z","timestamp":1723242674000},"score":1,"resource":{"primary":{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/10.1155\/2021\/6680509"}},"subtitle":[],"editor":[{"given":"Ning","family":"Cai","sequence":"additional","affiliation":[]}],"short-title":[],"issued":{"date-parts":[[2021,1]]},"references-count":35,"journal-issue":{"issue":"1","published-print":{"date-parts":[[2021,1]]}},"alternative-id":["10.1155\/2021\/6680509"],"URL":"https:\/\/doi.org\/10.1155\/2021\/6680509","archive":["Portico"],"relation":{},"ISSN":["1076-2787","1099-0526"],"issn-type":[{"type":"print","value":"1076-2787"},{"type":"electronic","value":"1099-0526"}],"subject":[],"published":{"date-parts":[[2021,1]]},"assertion":[{"value":"2020-12-06","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2021-01-24","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2021-02-08","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"6680509"}}