报告名称：PoseFlow: A Deep Motion Representation for Understanding Human Behaviors in Videos
Motion of the human body is the critical cue for understanding and characterizing human behavior in videos. Most existing approaches explore the motion cue using optical flows. However, optical flow usually contains motion on both the interested human bodies and the undesired background. This “noisy” motion representation makes it very challenging for pose estimation and action recognition in real scenarios. To address this issue, this talk presents a novel deep motion representation, called PoseFlow, which reveals human motion in videos while suppressing background and motion blur, and being robust to occlusion. For learning PoseFlow with mild computational cost, we propose a functionally structured spatial-temporal deep network, PoseFlow Net (PFN), to jointly solve the skeleton localization and matching problems of PoseFlow. Comprehensive experiments show that PFN outperforms the state-of-the-art deep flow estimation models in generating PoseFlow. Moreover, PoseFlow demonstrates its potential on improving two challenging tasks in human video analysis: pose estimation and action recognition.
Dingwen Zhang received the B.E. and Ph.D degrees from the Northwestern Polytechnical University, Xi’an, China, in 2012 and 2018, respectively. He is currently an associate professor in Xidian University. His research interests include computer vision and multimedia processing, especially on saliency detection, co-saliency detection, and weakly supervised learning.
报告名称：Person Re-identification by Multi-Channel Parts-Based CNN with Improved Triplet Loss Function
Person re-identification across cameras remains a very challenging problem, especially when there are no overlapping fields of view between cameras. In this talk, we present a novel multi-channel parts-based convolutional neural network (CNN) model under the triplet framework for person re-identification. Specifically, the proposed CNN model consists of multiple channels to jointly learn both the global full-body and local body-parts features of the input persons. The CNN model is trained by an improved triplet loss function that serves to pull the instances of the same person closer, and at the same time push the instances belonging to different persons farther from each other in the learned feature space. Extensive comparative evaluations demonstrate that our proposed method significantly outperforms many state-of-the-art approaches, including both traditional and deep network-based ones, on the challenging i-LIDS, VIPeR, PRID2011 and CUHK01 datasets.
De Cheng received the B.S. degree in automation control, in 2011, from Xi’an Jiaotong University, Xi’an, China, where he obtained his Ph.D. degree with the Institute of Artificial Intelligence and Robotics in 2017. He is now a faculty member in the Department of Computer Science and Technology. His research interests include pattern recognition and machine learning, specifically in the areas of person re-id, object detection and image classification.
报告名称：An object tracking algorithm combining spatial structure and motion continuity
To deal with the tracking drift problems caused by drastic object appearance change in complex scene, this talk proposes a robust tracking algorithm based on the sparse representation. It designs an optimized objective function with the spatial structure constraint. And then, with the Lagrange multiplier theory and accelerate proximal gradient approach, the coefficient of object template and candidate with the spatial information is obtained. In addition, the histogram intersection theory is exploited to computer the similarity between candidate and template. Finally, the template update scheme about when to manage updating and how to realize the strategy is presented, which combines the spatial structure information together with motion continuity and enables to tackle appearance change effectively. Experimental results on challenging benchmark datasets demonstrate that the novel algorithm performs favorable against several state-of-the-art methods.
Xiuhua Hu received the Ph.D. degree with the School of Automation,Northwestern Polytechnical University, Xi’an, China, in 2016. She is now a faculty member in Xi’an Technological University. Her research interests include pattern recognition and computer vision.