講座題目：Recent Developments in Motion Segmentation and Human Activity Recognition人類(lèi)活動(dòng)識(shí)別和運(yùn)動(dòng)分割的研究進(jìn)展
講座地點(diǎn)：計(jì)算機(jī)科學(xué)與軟件學(xué)院102會(huì)議室
講座時(shí)間：2015年4月24日（星期五）上午9:30
武慶明教授簡(jiǎn)介：
武慶明（Q.M. Jonathan Wu） 教授長(zhǎng)期從事計(jì)算機(jī)視覺(jué)，圖像處理，模式識(shí)別與智能系統(tǒng)的教學(xué)與研究工作，先后主持完成加拿大國(guó)家科學(xué)與工程研究項(xiàng)目(NSERC)，國(guó)際合作重大項(xiàng)目、加拿大國(guó)家重點(diǎn)基金項(xiàng)目，現(xiàn)任加拿大溫莎大學(xué)電子工程系教授，博士生導(dǎo)師，計(jì)算機(jī)視覺(jué)和傳感系統(tǒng)研究所主任,加拿大汽車(chē)電子和信息系統(tǒng)領(lǐng)域首席科學(xué)家，至今共培養(yǎng)博士、博士后30 余人，現(xiàn)任國(guó)際雜志《IEEE Transaction on Neural Networks and Learning Systems》, 《InternationalJournal of Robotics and Automation》與《Cognitive Computation》副主編，《IEEE Computational Intelligence Magazine》客座編委。曾任《IEEE Transaction on Systems, Man, and Cybernetics, Part A》與《International Journal of Control and Automation》編委。在過(guò)去的研究工作中，申請(qǐng)人對(duì)圖像實(shí)時(shí)分割、圖像壓縮與特征提取、圖像去噪與識(shí)別、三維重建等問(wèn)題進(jìn)行了深入研究，其研究成果均以通訊作者發(fā)表于《IEEE Transactions on Neural Networks and Learning Systems》、《IEEE Transactions on Image Processing》、《IEEE Transactions on Cybernetics》、《IEEE Transactions on Fuzzy Systems》、《International Journal of Computer Vision》、《Pattern Recognition》等國(guó)際相關(guān)學(xué)術(shù)期刊上，其中SCI 收錄論文100 多篇， IEEE Transactions 論文40 余篇，在圖像處理、智能信息處理、機(jī)器學(xué)習(xí)與模式識(shí)別領(lǐng)域，取得多項(xiàng)重大研究成果。武教授多次參加并任職包括ICCV、CVPR、ACCV、ICIP 等在內(nèi)三十多個(gè)國(guó)際會(huì)議的組織委員會(huì)主席和評(píng)審委員會(huì)委員。
講座內(nèi)容簡(jiǎn)介：
Motion segmentation in terms of dynamic textures, and human activity recognition are topics that have attracted growing attention in computer vision community. This talk is mainly concentrated on recent trends in dynamic texture segmentation, and human activity recognition. We first present two techniques for dynamic texture: a feature selection based dynamic mixture model for motion segmentation, and a linear-time video segmentation method which is scalable and temporally consistent for streaming videos. We then present a general framework to efficiently identify objects of interest in still images and later extend its application to human action recognition in videos. Such scheme can also be implemented in a situation where training data is coming in a serial mode and training needs to be performed in an incremental fashion. A brief overview of other related research activities in the presenter’s laboratory is also provided. Applications have been extended towards intelligent transportation systems, surveillance and security, face and gesture recognition, vision-guided robotics, and biomedical imaging, among others.
