导师姓名： 吴枫　 　
导师类型：
出生年份： 1969
职??称： 研究员
最高学历： 博士研究生 　
导师类别： 博士生导师
硕士学科： - 　 博士学科： 智能信息处理(081023) 　
工作单位： 微软亚洲研究院 办公电话： 86-10-58963119
通信地址： - 办公地点： -
电子邮箱： fengwu@microsoft.com 主页地址： http://research.microsoft.com/~fengwu/default.html
个人简历 吴枫博士，1992年毕业于西安电子科技大学电子工程专业，获工学学士学位。1994年考入哈尔滨工业大学模式识别与智能控制专业攻读硕士学位，从事语音识别研究，并于96年获得硕士学位。同年考入哈尔滨工业大学计算机应用专业攻读博士学位，从事视频编码研究，主要的研究方向有：多媒体信息编码和处理、多媒体通信、视觉信号处理，目前已经发表学术论文150余篇，其中包括近30篇相关研究领域的国际顶级期刊论文，并获得10余项美国专利。
Feng Wu received his B.S. degree in Electrical Engineering from Xidian Uni-versity in 1992. He received his M.S. and Ph.D. degrees in Computer Science from Harbin Institute of Technology in 1996 and 1999, respectively.
Wu joined Microsoft Research Asia, formerly named Microsoft Research China as an Associate Researcher in 1999. He has been a researcher with Microsoft Research Asia since 2001 and is now a Lead Researcher/Research Manager. His research interests include image and video representation, media compression and communication, as well as, computer vision and graphics. He has authored or co-authored over 150 papers published in journals like IEEE Transactions on Circuits and System for Video Technology, IEEE Transactions on Multimedia, IEEE Signal Processing Letters and some other International Conferences and Forums, e.g., ICIP, ICME and ISCAS. He has been an active contributor to ISO/MPEG and ITU-T standards. Some techniques have been adopted by MPEG-4 FGS, H.264/MPEG-4 AVC and H.264 SVC standard. Wu served as the chairman of China AVS video group in 2002~2004 and led the efforts on developing China AVS video standard 1.0. He has about 45 U.S. patents granted or pending in video and image coding.
Wu has been a senior member of IEEE. He serves as a reviewer for IEEE Transactions on Circuits and System for Video Technology, IEEE Transactions on Multimedia, IEEE Signal Processing Letters and several other International journals. He also serves as a member of the technical program committee of a number of International conferences, e.g., ICME 2006, PCS 2006, VCIP 2005.
主要研究方向及感兴趣的领域 1. Scalable Video Coding (SVC)
> PFGS
Progressive fine granularity scalable (PFGS) coding technology is an improvement of MPEG-4 FGS, where two motion compensations are used for each frame coding. Each frame at base layer is always predicted from the previous frame at base layer, whereas each frame at enhancement layer is predicted from the previous frame at either base layer or enhancement layer. Three coding modes are defined in PFGS to control the reference for prediction and reconstruction on each enhancement macroblock. Furthermore, a drifting model is proposed to estimate the drifting errors at encoder.
>3D subband video coding
In the 3D subband video coding, subband transform are applied in the horizontal, vertical and temporal directions, respectively. The resulted coefficients of each subband are scanned from one bit-plane to another and coded with either variable length table or arithmetic coding in a SNR scalable form. For the sake of high coding efficiency, motion alignment is incorporated into the temporal transform. Some techniques are developed to improve 3D subband video coding, such as Barbell lifting, 3D EBCOB, scalable motion vector coding, in-scale structure, etc.
2. Stream switching
Highly efficient adaptation on channel bandwidth is broadly required by streaming video over the Internet. Switching among non-scalable streams and/or scalable streams is a challenging topic because it may cause severe visual artifact and PSNR degradation due to the mismatch on the reconstructed references. Some techniques are developed and reported, such as improvements in the switching method of H.264/MPEG-4 AVC, switching among scalable streams, switching through distributed coding, etc.
3. Directional transform
2D DCT transform and 2D wavelet transform, which are extensively used in video and image coding, are implemented by two separable 1D transforms in horizontal and vertical directions. A serious drawback of these transforms is that they are ill suited to approximate image features with arbitrary orientation that is neither vertical nor horizontal. We have developed directional wavelet transform and DCT transform. Both of them are carried out in the direction of image edges and textures in a local window, and are not necessarily horizontal or vertical.
4. Vision-based image and video coding
Main-stream signal-processing-based compression schemes share a common architecture, namely transform followed by entropy coding, where only the statistical redundancy among pixels is considered as the adversary of coding. Through two-decade development, it has been becoming difficult to continuously improve the coding performance under such architecture. Based on newly developed vision technologies (e.g., inpainting and hallucination), we proposed a new image and video coding architecture to incorporate signal processing techniques and vision techniques together.
5. Distributed video coding (DVC)
The most attractive feature in DVC is that the complexity caused by motion estimation and compensation at the encoder can be shifted to the decoding side. The lightweight encoder becomes possible for portable and battery-powered devices. However, the distributed video coding also face many challenges, such as low coding efficiency caused by inaccurate side information, inefficiency entropy coding, how many WZ bits that should be sent to the decoder side. We are investigating the DVC schemes from these aspects and also exploiting kill application scenarios for DVC.
6. Multi-view and stereo video
With the developments on glass-free display devices of multi-view and stereo video, applications of immersive media get more and more attention, such as 3D TV and immersive conference. Our researches involve two aspects. One is how to efficiently represent of multi-view video in the compressed way because the correlation among views provides a bigger space to exploit in compression. Another is how to generate multi-video and stereo content from existing video and image.
7. Intermedia
There are two different approaches for network and device adaptation in video coding: scalable video coding and transcoding. Streams generated by scalable video coding are very easy to modify its resolution, frame rate and bit rate by the truncating process. However, scalable video coding suffers from coding efficiency degradation. Transcoding can provide high performance when modifying streams from a combination of resolution, frame rate and bit rate to another. But the transcoding process is of high computational intensity. The Intermedia tries to generate a representation of source video and/or compressed stream to facilitate the network and device adaptation in an easy and efficient way.
研究成果与著作个人代表作获奖情况 -
目前承担的科研项目及教学情况 CONTACT INFORMATION
MICROSOFT RESEARCH CHINA
5F Sigma, No49 Zhichun Rd
Haidian, Beijing, 100080, China
Email: fengwu@microsoft.com
Phone: (86-10) 58963119
Fax: (86-10) 88097306
有关招生的附加说明 校外兼职导师 现为电子工程学院智能信息处理研究所客座教授
信息来源：
http://today.hit.edu.cn/articles/2004/6-11/17168.htm
http://research.microsoft.com/~fengwu/default.html
http://icme07.ustc.edu.cn/Speakers.htm
信息来源声明 信息来源声明：本导师简介来源于西安电子科技大学 电子工程学院 网站，转载时请保留此声明
http://see.xidian.edu.cn(教育网，网通用户) http://see.xidian.sn.cn（电信ADSL等用户）
新导师不断加入以及导师简介不断更新！
请访问电院网站查看最新的 导师简介以及研究生招生信息
电院首页——>研究生培养——>导师介绍
如果有问题，请反馈在留言簿上。如果需要申请个人主页空间 以及更换照片，请联系管理员 seewebmaster@163.com。导师的信息仅供考生报考所用，不得用于商业等目的。