报告主题：基于立体对比度和立体聚焦的立体显著度检测
主讲人：A/ Prof. Jian Zhang, 澳大利亚悉尼科技大学（UTS）
时间：2015年1月7日14:15
地点：延长校区行健楼1018室
演讲者简介：
Jian Zhang received the B.Sc. degree from East Normal University, China in 1982; the M.Sc. degree in Computer Science from Flinders University, Australia in 1994; and the Ph.D. degree in Electrical Engineering from the University of New South Wales, Australia in 1999.
From 1997 to 2003, A/ProfZhang was with the Visual Information Processing Laboratory, Motorola Labs, Sydney, as a Senior Research Engineer, and later became a Principal Research Engineer and a Foundation Manager with the Visual Communications Research Team. From 2004 to July 2011, he was a Principal Researcher and a Project Leader with National ICT Australia, Sydney, and a Conjoint Associate Professor with the School of Computer Science and Engineering, UNSW. He is currently an Associate Professor with the Advanced Analytics Institute, School of software, Faculty of engineering and Information Technology, University of Technology Sydney, Sydney. A/Prof Zhang’s research interests include multimedia signal processing, computer vision, pattern recognition, visual information mining, human-computer interaction and intelligent video surveillance systems.Apart from more than 100 paper publications, book chapters, patents and technical reports from his research output, he was co-author of more than ten patents filed in US, UK, Japan and Australia including six issued US patents.
Dr. Zhang is an IEEE Senior Member, Associated Editors for IEEE Transactions on Circuits and Systems for Video Technology (T-CSVT) and EURASIP Journal on Image and Video Processing. As a general co-chair, he has hosted the 2012 IEEE International Conference on Multimedia and Expo in Melbourne, Australia. As a Technical Program Co-chair, he chaired the IEEE International Workshop on Multimedia Signal Processing (MMSP) 2008 and IEEE Visual Communications and Image Processing (VCIP), 2014.
讲座摘要：
Saliency detection is an important branch in computer vision, which is widely used in computer vision field. Currently, there are many different saliency detection algorithms proposed for 2D images/video. With the rapid development of stereo display technology, there are more and more applications emerging for 3D image or videos, which increase the demands of the saliency detection models for 3D visual content. Comparing with 2D visual saliency model, stereo saliency detection model needs to take depth (disparity) factor into consideration. Different from color factors, the depth factors have its unique characteristics which make the 3D saliency detection model as an ongoing research question in human stereo vision. In this talk, we analyse two characteristics of human stereo vision. Based on different disparities, the objects in stereo image or videos have pop-out or deep-in effect which impacts on human attention (attractive or not attractive). The second is comfortable zone. In order to avoid 3D fatigue, a good stereo image might distribute the disparity of some important object in a special zone known as comfortable zone. Having analysed the two characteristics, we propose stereo saliency detection based on stereo contrast and stereo focus. Stereo contrast is used to measure stereo saliency based on pop-out effect and color contrast. Stereo focus is used to describe stereo focus degree of binocular vision based on comfortable zone and monocular focus. After the stereo saliency detection, we then cluster the stereo contrast and stereo focus on their histograms. Finally, we integration the stereo contrast and its focus by applying the multi-scale detections and clustering processes. Experimental results on a recent eye-tracking database have shown that the proposed method outperform other existing methods in stereo saliency detection for 3D images.