{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,10]],"date-time":"2026-06-10T02:08:05Z","timestamp":1781057285862,"version":"3.54.1"},"reference-count":92,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","issue":"10","license":[{"start":{"date-parts":[[2018,10,1]],"date-time":"2018-10-01T00:00:00Z","timestamp":1538352000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"}],"funder":[{"DOI":"10.13039\/501100012234","name":"Shenzhen Peacock Plan","doi-asserted-by":"crossref","award":["20130408-183003656"],"award-info":[{"award-number":["20130408-183003656"]}],"id":[{"id":"10.13039\/501100012234","id-type":"DOI","asserted-by":"crossref"}]},{"name":"Shenzhen Key Laboratory for Intelligent Multimedia and Virtual Reality","award":["ZDSYS201703031405467"],"award-info":[{"award-number":["ZDSYS201703031405467"]}]},{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"crossref","award":["U1613209"],"award-info":[{"award-number":["U1613209"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"crossref"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. Multimedia"],"published-print":{"date-parts":[[2018,10]]},"DOI":"10.1109\/tmm.2018.2855081","type":"journal-article","created":{"date-parts":[[2018,7,11]],"date-time":"2018-07-11T18:47:39Z","timestamp":1531334859000},"page":"2578-2592","source":"Crossref","is-referenced-by-count":26,"title":["Multiscale Deep Alternative Neural Network for Large-Scale Video Classification"],"prefix":"10.1109","volume":"20","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-9464-4426","authenticated-orcid":false,"given":"Jinzhuo","family":"Wang","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2664-4413","authenticated-orcid":false,"given":"Wenmin","family":"Wang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Wen","family":"Gao","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref73","first-page":"646","article-title":"Deep networks with stochastic depth","author":"huang","year":"0","journal-title":"Proc Eur Conf Comput Vis"},{"key":"ref72","article-title":"Distilling\n the knowledge in a neural network","author":"hinton","year":"2015"},{"key":"ref71","first-page":"1929","article-title":"Dropout: A simple way to prevent neural networks from\n overfitting","volume":"15","author":"srivastava","year":"2014","journal-title":"J Mach Learn Res"},{"key":"ref70","first-page":"2261","article-title":"Densely connected\n convolutional networks","author":"huang","year":"0","journal-title":"Proc IEEE Conf Comput Vis Pattern Recognit"},{"key":"ref76","first-page":"4281","article-title":"Sobolev training for neural networks","author":"czarnecki","year":"0","journal-title":"Proc 31st Int Conf Neural Inf Process Syst"},{"key":"ref77","first-page":"191","article-title":"Electrical activity of muscles of the trunk during walking","volume":"111","author":"waters","year":"1972","journal-title":"J Anatomy"},{"key":"ref74","article-title":"Very deep convolutional networks for large-scale image recognition","author":"simonyan","year":"2014"},{"key":"ref39","doi-asserted-by":"publisher","DOI":"10.1007\/s11263-005-1838-7"},{"key":"ref75","first-page":"770","article-title":"Deep residual learning for image recognition","author":"he","year":"0","journal-title":"Proc IEEE Conf Comput Vis Pattern Recognit"},{"key":"ref38","first-page":"3185","article-title":"Action\n recognition with multiscale spatio-temporal contexts","author":"wang","year":"0","journal-title":"Proc IEEE Conf Comput Vis Pattern Recognit"},{"key":"ref78","first-page":"1458","article-title":"The pyramid match kernel: Discriminative classification with sets of image features","volume":"2","author":"grauman","year":"0","journal-title":"Proc Int Conf Comput Vis"},{"key":"ref79","first-page":"2169","article-title":"Beyond bags\n of features: Spatial pyramid matching for recognizing natural scene categories","volume":"2","author":"lazebnik","year":"0","journal-title":"Proc IEEE Conf Comput Vis Pattern Recognit"},{"key":"ref33","first-page":"32","article-title":"Recognizing\n human actions: A local SVM approach","volume":"3","author":"sch\u00fcldt","year":"0","journal-title":"Proc 17th Int Conf Pattern Recognit"},{"key":"ref32","first-page":"469","article-title":"Analyzing and recognizing walking figures in XYT","author":"niyogi","year":"0","journal-title":"Proc IEEE Conf Comput Vis Pattern Recognit"},{"key":"ref31","doi-asserted-by":"publisher","DOI":"10.1109\/TMM.2016.2635030"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.1109\/TMM.2017.2674622"},{"key":"ref37","first-page":"1097","article-title":"ImageNet\n classification with deep convolutional neural networks","author":"krizhevsky","year":"0","journal-title":"Proc 25th Int Conf Neural Inf Process Syst"},{"key":"ref36","first-page":"124.1","article-title":"Evaluation of local spatio-temporal features for action\n recognition","author":"wang","year":"0","journal-title":"Proc Brit Mach Vis Conf"},{"key":"ref35","first-page":"143","article-title":"Improving the fisher kernel for large-scale image classification","author":"perronnin","year":"0","journal-title":"Proc Eur Conf Comput Vis"},{"key":"ref34","first-page":"1","article-title":"Learning realistic\n human actions from movies","author":"laptev","year":"0","journal-title":"Proc IEEE Conf Comput Vis Pattern Recognit"},{"key":"ref60","first-page":"18","article-title":"Convolutional recurrent neural networks: Learning spatial\n dependencies for image representation","author":"zuo","year":"0","journal-title":"Proc IEEE Conf Comput Vis Pattern Recognit Workshop"},{"key":"ref62","article-title":"Understanding deep architectures using a\n recursive convolutional network","author":"eigen","year":"2013"},{"key":"ref61","first-page":"656","article-title":"Convolutional-recursive deep learning for 3D object classification","author":"socher","year":"0","journal-title":"Proc 25th Int Conf Neural Inf Process Syst"},{"key":"ref63","first-page":"625","article-title":"Why does unsupervised pre-training help deep\n learning","volume":"11","author":"erhan","year":"2010","journal-title":"J Mach Learning Res"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.1109\/TMM.2016.2535357"},{"key":"ref64","first-page":"2377","article-title":"Training very deep networks","author":"srivastava","year":"0","journal-title":"Proc 28th Int Conf Neural Inf Process Syst"},{"key":"ref27","doi-asserted-by":"publisher","DOI":"10.1109\/TMM.2013.2293060"},{"key":"ref65","first-page":"770","article-title":"Deep residual learning for image recognition","author":"he","year":"0","journal-title":"Proc IEEE Conf Comput Vis Pattern Recognit"},{"key":"ref66","first-page":"448","article-title":"Batch normalization: Accelerating deep network training by reducing internal covariate shift","author":"ioffe","year":"0","journal-title":"Proc Int Conf Mach Learn"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.1109\/TMM.2017.2666540"},{"key":"ref67","article-title":"Object detectors emerge in deep scene CNNs","author":"zhou","year":"0"},{"key":"ref68","first-page":"562","article-title":"Deeply-supervised nets","author":"lee","year":"0","journal-title":"Proc Int Conf Artif Intell Statist"},{"key":"ref69","first-page":"1","article-title":"Going deeper with convolutions","author":"szegedy","year":"0","journal-title":"Proc IEEE Conf Comput Vis Pattern Recognit"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.1145\/3150226"},{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1109\/TMM.2017.2760623"},{"key":"ref20","first-page":"811","article-title":"Deep alternative neural network: Exploring contexts as early as\n possible for action recognition","author":"wang","year":"0","journal-title":"Proc 30th Int Conf Neural Inf Process Syst"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1109\/93.311653"},{"key":"ref21","first-page":"937","article-title":"Convolutional neural networks with intra-layer recurrent connections for scene labeling","author":"liang","year":"0","journal-title":"Proc 28th Int Conf Neural Inf Process Syst"},{"key":"ref24","doi-asserted-by":"publisher","DOI":"10.1006\/cviu.1998.0711"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.1145\/211430.211433"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.1109\/TMM.2011.2176110"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.1109\/TMM.2011.2128301"},{"key":"ref50","first-page":"3097","article-title":"Spatio-temporal vector of locally max\n pooled features for action recognition in videos","author":"duta","year":"0","journal-title":"Proc IEEE Conf Comput Vis Pattern Recognit"},{"key":"ref51","first-page":"2097","article-title":"Spatiotemporal pyramid network for video\n action recognition","author":"wang","year":"0","journal-title":"Proc IEEE Conf Comput Vis Pattern Recognit"},{"key":"ref92","first-page":"618","article-title":"Grad-CAM: Visual explanations from deep\n networks via gradient-based localization","author":"selvaraju","year":"0","journal-title":"Proc Int Conf Comput Vis"},{"key":"ref91","first-page":"716","article-title":"Learning long-term dependencies for action recognition with a biologically-inspired deep network","author":"shi","year":"0","journal-title":"Proc Int Conf Comput Vis"},{"key":"ref90","first-page":"4757","article-title":"Procedural generation of\n videos to train deep action recognition networks","author":"de souza","year":"0","journal-title":"Proc IEEE Conf Comput Vis Pattern Recognit"},{"key":"ref59","first-page":"4694","article-title":"Beyond short snippets: Deep networks for video\n classification","author":"ng","year":"0","journal-title":"Proc IEEE Conf Comput Vis Pattern Recognit"},{"key":"ref58","first-page":"2625","article-title":"Long-term recurrent convolutional networks for\n visual recognition and description","author":"donahue","year":"0","journal-title":"Proc Int Conf Comput Vis"},{"key":"ref57","first-page":"6402","article-title":"UntrimmedNets for weakly supervised action\n recognition and detection","author":"wang","year":"0","journal-title":"Proc IEEE Conf Comput Vis Pattern Recognit"},{"key":"ref56","first-page":"668","article-title":"Depth2action: Exploring embedded depth for large-scale action recognition","author":"zhu","year":"0","journal-title":"Proc Eur Conf Comput Vis"},{"key":"ref55","first-page":"2166","article-title":"Lattice long short-term memory for human action recognition","author":"sun","year":"0","journal-title":"Proc Int Conf Comput Vis"},{"key":"ref54","first-page":"1991","article-title":"A key volume mining deep framework for action recognition","author":"zhu","year":"0","journal-title":"Proc IEEE Conf Comput Vis Pattern Recognit"},{"key":"ref53","first-page":"1933","article-title":"Convolutional two-stream network fusion for video action recognition","author":"feichtenhofer","year":"0","journal-title":"Proc IEEE Conf Comput Vis Pattern Recognit"},{"key":"ref52","doi-asserted-by":"publisher","DOI":"10.1007\/s11263-015-0816-y"},{"key":"ref10","first-page":"568","article-title":"Two-stream convolutional networks for action recognition in videos","author":"simonyan","year":"0","journal-title":"Proc Conf Neural Inf Process Syst"},{"key":"ref11","first-page":"4305","article-title":"Action\n recognition with trajectory-pooled deep-convolutional descriptors","author":"wang","year":"0","journal-title":"Proc IEEE Conf Comput Vis Pattern Recognit"},{"key":"ref40","first-page":"650","article-title":"An efficient dense and scale-invariant spatio-temporal interest point detector","author":"willems","year":"0","journal-title":"Proc Eur Conf Comput Vis"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2012.59"},{"key":"ref13","first-page":"4489","article-title":"Learning spatiotemporal features with 3D convolutional networks","author":"tran","year":"0","journal-title":"Proc Int Conf Comput Vis"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2017.2712608"},{"key":"ref15","first-page":"20","article-title":"Temporal segment networks: Towards good\n practices for deep action recognition","author":"wang","year":"0","journal-title":"Proc Eur Conf Comput Vis"},{"key":"ref82","first-page":"556","article-title":"Semantic concept detection using weighted discretization multiple correspondence analysis for\n disaster information management","author":"pouyanfar","year":"0","journal-title":"Proc IEEE Int Conf Inf Reuse Integration"},{"key":"ref16","first-page":"204","article-title":"Beyond gaussian pyramid: Multi-skip feature stacking for action recognition","author":"lan","year":"0","journal-title":"Proc IEEE Conf Comput Vis Pattern Recognit"},{"key":"ref81","doi-asserted-by":"crossref","first-page":"1550","DOI":"10.1109\/5.58337","article-title":"Backpropagation through time: What it does and how to do it","volume":"78","author":"werbos","year":"0","journal-title":"Proc IEEE"},{"key":"ref17","first-page":"581","article-title":"Action recognition with stacked fisher\n vectors","author":"peng","year":"0","journal-title":"Proc Eur Conf Comput Vis"},{"key":"ref84","first-page":"2593","article-title":"Efficient feature extraction, encoding and classification for action recognition","author":"kantorov","year":"0","journal-title":"Proc IEEE Conf Comput Vis Pattern Recognit"},{"key":"ref18","first-page":"3551","article-title":"Action recognition with improved trajectories","author":"wang","year":"0","journal-title":"Proc Int Conf Comput Vis"},{"key":"ref83","article-title":"Torch7: A MATLAB-like environment for machine learning","author":"collobert","year":"0","journal-title":"NIPS Workshop BigLearn"},{"key":"ref19","first-page":"2447","article-title":"Recursive\n context propagation network for semantic scene labeling","author":"sharma","year":"0","journal-title":"Proc Int Conf Neural Inf Process"},{"key":"ref80","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2015.2389824"},{"key":"ref89","first-page":"4724","article-title":"Quo vadis, action recognition? A new model and the kinetics dataset","author":"carreira","year":"0","journal-title":"Proc IEEE Conf Comput Vis Pattern Recognit"},{"key":"ref4","first-page":"791","article-title":"Multi-stream multi-class fusion of deep networks for video classification","author":"wu","year":"0","journal-title":"Proc ACM Multimedia Conf"},{"key":"ref3","doi-asserted-by":"crossref","first-page":"109","DOI":"10.1016\/j.cviu.2016.03.013","article-title":"Bag of visual words and fusion methods for\n action recognition: Comprehensive study and good practice","volume":"150","author":"peng","year":"2016","journal-title":"Comput Vis Image Understanding"},{"key":"ref6","first-page":"2556","article-title":"HMDB: A large video database for human motion recognition","author":"kuehne","year":"0","journal-title":"Proc Int Conf Comput Vis"},{"key":"ref5","article-title":"UCF101: A\n dataset of 101 human actions classes from videos in the wild","author":"soomro","year":"2012"},{"key":"ref85","first-page":"25","article-title":"High accuracy optical flow estimation\n based on a theory for warping","author":"brox","year":"0","journal-title":"Proc Eur Conf Comput Vis"},{"key":"ref8","first-page":"2980","article-title":"Mask R-CNN","author":"he","year":"0","journal-title":"Proc Int Conf Comput Vis"},{"key":"ref86","first-page":"214","article-title":"A duality\n based approach for realtime TV-L1 optical flow","author":"zach","year":"0","journal-title":"Proc 29th DAGM Conf Pattern Recognit"},{"key":"ref7","first-page":"961","article-title":"ActivityNet: A large-scale video\n benchmark for human activity understanding","author":"heilbron","year":"0","journal-title":"Proc IEEE Conf Comput Vis Pattern Recognit"},{"key":"ref49","first-page":"1219","article-title":"Deep local video feature for action recognition","author":"lan","year":"0","journal-title":"Proc IEEE Conf Comput Vis Pattern Recognit Workshop"},{"key":"ref87","article-title":"Hidden two-stream convolutional networks for\n action recognition","author":"zhu","year":"2017"},{"key":"ref88","doi-asserted-by":"publisher","DOI":"10.1007\/s11263-015-0859-0"},{"key":"ref9","first-page":"1725","article-title":"Large-scale video classification with\n convolutional neural networks","author":"karpathy","year":"0","journal-title":"Proc IEEE Conf Comput Vis Pattern Recognit"},{"key":"ref46","first-page":"5699","article-title":"AdaScan: Adaptive scan pooling in deep convolutional\n neural networks for human action recognition in videos","author":"kar","year":"0","journal-title":"Proc IEEE Conf Comput Vis Pattern Recognit"},{"key":"ref45","first-page":"565","article-title":"Video action\n detection with relational dynamic-poselets","author":"wang","year":"0","journal-title":"Proc Eur Conf Comput Vis"},{"key":"ref48","first-page":"7445","article-title":"Spatiotemporal multiplier networks for video action recognition","author":"feichtenhofer","year":"0","journal-title":"Proc IEEE Conf Comput Vis Pattern Recognit"},{"key":"ref47","first-page":"5534","article-title":"Learning spatio-temporal\n representation with pseudo-3D residual networks","author":"qiu","year":"0","journal-title":"Proc Int Conf Comput Vis"},{"key":"ref42","first-page":"357","article-title":"A\n 3-dimensional sift descriptor and its application to action recognition","author":"scovanner","year":"0","journal-title":"Proc 15th ACM Int Conf Multimedia"},{"key":"ref41","first-page":"275:1","article-title":"A spatio-temporal descriptor based on 3D-gradients","author":"klaser","year":"0","journal-title":"Proc 19th Brit Mach Vis Conf"},{"key":"ref44","first-page":"1234","article-title":"Action bank: A high-level representation of activity in video","author":"sadanand","year":"0","journal-title":"Proc IEEE Conf Comput Vis Pattern Recognit"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.1007\/s11263-012-0594-8"}],"container-title":["IEEE Transactions on Multimedia"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx7\/6046\/8466674\/08409973.pdf?arnumber=8409973","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,1,26]],"date-time":"2022-01-26T09:06:41Z","timestamp":1643188001000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/8409973\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2018,10]]},"references-count":92,"journal-issue":{"issue":"10"},"URL":"https:\/\/doi.org\/10.1109\/tmm.2018.2855081","relation":{},"ISSN":["1520-9210","1941-0077"],"issn-type":[{"value":"1520-9210","type":"print"},{"value":"1941-0077","type":"electronic"}],"subject":[],"published":{"date-parts":[[2018,10]]}}}