{"668862":{"#nid":"668862","#data":{"type":"news","title":"Researchers Awarded EURASIP 2023 Best Paper Prize for Innovative Video Action Recognition Techniques","body":[{"value":"\u003Cp\u003EA team of researchers, led by Professor \u003Ca href=\u0022https:\/\/ece.gatech.edu\/directory\/ghassan-alregib\u0022\u003EGhassan AlRegib\u003C\/a\u003E in the \u003Ca href=\u0022https:\/\/ece.gatech.edu\/\u0022\u003ESchool of Electrical and Computer Engineering\u003C\/a\u003E, has been awarded the prestigious \u003Ca href=\u0022https:\/\/eurasip.org\/\u0022\u003EEuropean Association For Signal Processing (EURASIP)\u003C\/a\u003E 2023 \u003Ca href=\u0022https:\/\/eurasip.org\/best-paper-awards\/\u0022\u003EBest Paper Award\u003C\/a\u003E for their outstanding contribution to the Image Communication journal.\u003C\/p\u003E\u003Cp\u003EThe award-winning paper, titled \u0022\u003Ca href=\u0022https:\/\/arxiv.org\/pdf\/1703.10667.pdf\u0022\u003ETS-LSTM and temporal-inception: Exploiting spatiotemporal dynamics for activity recognition\u003C\/a\u003E,\u0022 was authored in 2019 by AlRegib and Assistant Professor \u003Ca href=\u0022https:\/\/www.cc.gatech.edu\/people\/zsolt-kira\u0022\u003EZsolt Kira\u003C\/a\u003E in the \u003Ca href=\u0022https:\/\/www.cc.gatech.edu\/\u0022\u003ECollege of Computing\u003C\/a\u003E, as well as then-Ph.D. candidates \u003Ca href=\u0022https:\/\/chihyaoma.github.io\/\u0022\u003EChih-Yao (Kevin) Ma\u003C\/a\u003E (now a staff research scientist at Meta) and \u003Ca href=\u0022https:\/\/minhungchen.netlify.app\/\u0022\u003EMin-Hung (Steve) Chen\u003C\/a\u003E (now a senior research scientist at NVIDIA). the work was completed in AlRegib\u2019s \u003Ca href=\u0022https:\/\/ghassanalregib.info\/\u0022\u003EOmni Lab for Intelligent Visual Engineering and Science (OLIVES)\u003C\/a\u003E.\u003C\/p\u003E\u003Cp\u003EThe research explores different methods to improve action recognition in videos using deep neural networks, specifically two-stream Convolutional Neural Networks (ConvNets). It investigates how to better capture the spatiotemporal dynamics within video sequences. The study proposes and evaluates two new network architectures: temporal segment RNN and Inception-style Temporal-ConvNet. These architectures aim to effectively integrate spatiotemporal information and enhance the overall performance of action recognition systems. The research also identifies specific limitations of each method, which could guide future work in this area.\u003C\/p\u003E\u003Cp\u003EThe EURASIP Best Paper Award is bestowed upon exceptional papers published in journals sponsored by EURASIP within the past four years. This ensures that the most influential and vital ideas in the field are honored.\u003C\/p\u003E\u003Cp\u003EThe Georgia Tech team will be honored at the \u003Ca href=\u0022https:\/\/eusipco2023.org\/\u0022\u003EEuropean Signal Processing Conference (EUSIPCO)\u003C\/a\u003E held in Helsinki, Finland, from September 4-8. EUSIPCO offers a comprehensive technical program addressing the latest developments in signal processing research and technology, along with its diverse applications.\u003C\/p\u003E\u003Cp\u003EEURASIP, founded in 1978, has been a crucial platform for researchers, providing an academic and professional forum to disseminate and discuss all aspects of signal processing, fostering growth and collaboration within the community.\u003Cbr\u003E\u0026nbsp;\u003C\/p\u003E\u003Cp\u003E\u003Cem\u003E\u003Cstrong\u003ECaption for photo header:\u003C\/strong\u003E Overview of the proposed framework. Spatial and temporal features were extracted from a two-stream ConvNet using ResNet-101 pre-trained on ImageNet, and fine-tuned for single-frame activity prediction. Spatial and temporal features are concatenated and temporally-constructed into feature matrices. The constructed feature matrices are then used as input to both of our proposed methods: Temporal Segment LSTM (TS-LSTM) and Temporal-Inception.\u003C\/em\u003E\u003C\/p\u003E","summary":"","format":"limited_html"}],"field_subtitle":"","field_summary":[{"value":"\u003Cp\u003EThe research explores different methods to improve action recognition in videos using deep neural networks, specifically two-stream Convolutional Neural Networks (ConvNets).\u0026nbsp;\u003C\/p\u003E\r\n","format":"limited_html"}],"field_summary_sentence":[{"value":"Professor AlRegib and team received the EURASIP 2023 Best Paper Award for their novel methods to improve action recognition in videos using deep neural networks."}],"uid":"36172","created_gmt":"2023-08-09 14:40:15","changed_gmt":"2024-06-24 13:01:42","author":"dwatson71","boilerplate_text":"","field_publication":"","field_article_url":"","dateline":{"date":"2023-08-09T00:00:00-04:00","iso_date":"2023-08-09T00:00:00-04:00","tz":"America\/New_York"},"extras":[],"hg_media":{"671352":{"id":"671352","type":"image","title":"ConvNets Framwork.jpg","body":"\u003Cp\u003E\u003Cspan\u003E\u003Cspan\u003E\u003Cspan\u003E\u003Cspan\u003E\u003Cspan\u003E\u003Cspan\u003E\u003Cspan\u003EOverview of the proposed framework. Spatial and temporal features were extracted from a two-stream ConvNet using ResNet-101 pre-trained on ImageNet, and fine-tuned for single-frame activity prediction. Spatial and temporal features are concatenated and temporally-constructed into feature matrices. The constructed feature matrices are then used as input to both of our proposed methods: Temporal Segment LSTM (TS-LSTM) and Temporal-Inception. \u003C\/span\u003E\u003C\/span\u003E\u003C\/span\u003E\u003C\/span\u003E\u003C\/span\u003E\u003C\/span\u003E\u003C\/span\u003E\u003C\/p\u003E\r\n","created":"1691592031","gmt_created":"2023-08-09 14:40:31","changed":"1691592031","gmt_changed":"2023-08-09 14:40:31","alt":"A graphic of the proposed framework with Temporal Segment LSTM (TS-LSTM) and Temporal-Inception.","file":{"fid":"254385","name":"ConvNets Framwork.jpg","image_path":"\/sites\/default\/files\/2023\/08\/09\/ConvNets%20Framwork.jpg","image_full_path":"http:\/\/hg.gatech.edu\/\/sites\/default\/files\/2023\/08\/09\/ConvNets%20Framwork.jpg","mime":"image\/jpeg","size":422146,"path_740":"http:\/\/hg.gatech.edu\/sites\/default\/files\/styles\/740xx_scale\/public\/2023\/08\/09\/ConvNets%20Framwork.jpg?itok=MLZjuAsY"}}},"media_ids":["671352"],"groups":[{"id":"1255","name":"School of Electrical and Computer Engineering"}],"categories":[],"keywords":[{"id":"44681","name":"Ghassan AlRegib"},{"id":"192932","name":"European Association For Signal Processing"},{"id":"192933","name":"TS-LSTM and temporal-inception"},{"id":"654","name":"College of Computing"},{"id":"192934","name":"Chih-Yao Ma"},{"id":"192935","name":"Min-Hung Chen"},{"id":"178069","name":"Omni Lab for Intelligent Visual Engineering and Science"},{"id":"66891","name":"Georgia Tech School of Electrical and Computer Engineering"},{"id":"192936","name":"European Signal Processing Conference"},{"id":"192937","name":"Convolutional Neural Networks"}],"core_research_areas":[{"id":"39431","name":"Data Engineering and Science"}],"news_room_topics":[],"event_categories":[],"invited_audience":[],"affiliations":[],"classification":[],"areas_of_expertise":[],"news_and_recent_appearances":[],"phone":[],"contact":[{"value":"\u003Cp\u003EDan Watson\u003C\/p\u003E","format":"limited_html"}],"email":["dwatson@ece.gatech.edu"],"slides":[],"orientation":[],"userdata":""}}}