About AMP Lab        Projects        Downloads        Publications        People        Links

Group Member

	Xiaoming Liu PhD. Student Personal Homepage: http://www.andrew.cmu.edu/~xiaoming Email: xiaoming@andrew.cmu.edu
	Office: Porter Hall B41 Lab: Porter Hall B6 Phone: 412-268-7100 Fax: 412-268-3890	Mailing Address: Department of ECE, Carnegie Mellon University, 5000 Forbes Avenue, Pittsburgh, PA 15213-3890

[Research Interests]        [Project]      [Publications] 

• Video based face recognition
  Although face recognition has been an active research topic for decades, the traditional recognition algorithms are all based on static images. In this project, we would like to propose algorithms for video based face recognition because it has superior advantages over the image-based recognition. Such as, the temporal information of faces can be utilized to facilitate the recognition task; and further, a better modeling technique can be applied to the video sequence of human faces.

• Model updating for pattern recognition
  In patter recognition and computer vision, appearance modeling is an important way for object tracking and recognition. While many modeling methods have been proposed, few of them were dealing with the updating of the model, which is very useful because of the lacking training data and always under-going changing of the object's appearance. In this work we propose to update the eigenspace model according the changing statistics of the samples being observed. 

• Face Authentication
  With electronic communication starting to dominate large areas of every day life, a safe method of verifying user is essential to prevent data misuse. Because of the appearance variations in faces, the authentication of persons based on human faces is a complex task with high performance and robustness requirements when used in practical applications. We are interested in different ways to deal with the variations in face images, such as expression, registration error, illumination and poses.

• Video Based Human Animation
  My earlier work concentrate on the analysis of human motion in realistic scene and its application in human animation. Human animation is a challenging domain in computer animation. To arm at many shortcomings of conventional techniques, we propose a new video based human animation technique. The key idea of this approach is to recover the 3D human motion information from a monocular  2D image sequence.  

• Multimedia information retrieval
   With the fast development of Internet and digital technology, more and more data are readily available on the web. How to extract the needed information from the deep sea of web has become a critical issue, which is identified as web-based Multimedia Information Retrieval (MIR). The initiative of this research is to explore ways of intelligent MIR. In Zhejiang University, I concentrated my work on the video similarity model, and using semantic template to support image retrieval.

Top of this page

Video Based Face Recognition Video based face recognition is studied in this project. For making use of the temporal information, we propose to apply temporal and adaptive Hidden Markov Model (HMM) on face sequence for recognition. For having better modeling of different variations of human faces, a face mosaic model is obtained from training face sequences and used for recognition. For more details on the project please see the Video based face recognition project page.   In patter recognition and computer vision, appearance modeling is an important way for object tracking and recognition. Considering the appearance of the object is always changing, to learn the changing and build an adaptive model are very critical for many applications. In this project, we propose to update the eigenspace model in order to adaptively learn the changing appearance of an object, and use it in different applications. For more details on the project please see the Model updating project page.   Eigenflow Based Face Authentication Currently the computer only uses password to authenticate the user, we hope  computer can intelligently implement this task by the biometrics approach, i.e. let the face, the voice and the fingerprint of the user do authentication.  The face authentication is a challenging research topic since the human face can always generate significant variations in appearance because of facial expressions, rotation, scale, shift, lighting condition, etc. Here we propose a new approach to perform face authentication tolerant to facial expression variations and registration error. For more details on the project please see the Face authentication project page.   Video Based Human Animation  In this project, when given a clip of video, we can track human joint in image sequence, then construct corresponding three-dimension human motion skeleton sequence under the perspective projection by use of the camera calibration and human anatomy knowledge, and finally establish a motion library by annotating multiform motion attributes automatically, which can be browsed and queried by the animator. This approach has the characteristic of rich source material, low computing cost, efficient production, and realistic animation result. We demonstrate this on several video clips of people doing full body movements, and visualize the results in re-animating a 3D human skeleton model.

Top of this page

Journals Articles:

• Xiaoming Liu, Tsuhan Chen and Susan M. Thornton, Eigenspace Updating for Non-Stationary Process and Its Application to Face Recognition. Accepted for publication on Pattern Recognition, Special issue on Kernel and Subspace Methods for Computer Vision. September, 2002.

• Xiaoming Liu, Tsuhan Chen and B.V.K. Vijaya Kumar, Face Authentication for Multiple Subjects Using Eigenflow. Pattern Recognition, special issue on Biometric,Volume 36, Issue 2, February 2003, pp. 313-328.

• Liu xiaoming, Zhuang yueting and Pan yunhe. Model Based Human Motion Tracking. Journal of Computer Research & Development. Vol.36,No.10, Oct.1999.(in Chinese)

• Zhuang yueting, Liu xiaoming. Pan yunhe and Yang jun. Human Three Dimension Motion Skeleton Reconstruction of Motion Image Sequence. Accepted for publication in Journal of Computer-aided Design & Computer Graphics. March,1999.(in Chinese)

• Liu xiaoming, Zhuang yueting and Pan yunhe. A Video Based Human Skeleton Extraction Technique For Animation. Accepted for publication in Journal of Computer Research & Development. Nov. 1999(in Chinese)

• Zhuang yueting, Liu xiaoming, Chen chun, Zhang chenyu and Lu weilin. GDP: The Design and Implementation of Object Oriented and Software Components Techniques. Accepted for publication in Computer Science. June,1999(in Chinese)  

          Conferences Papers:

• Xiaoming Liu and Tsuhan Chen, Geometry-assisted Statistical Modeling for Face Mosaicing, submitted to ICIP 2003.

• Xiaoming Liu, Tsuhan Chen, Video-Based Face Recognition Using Adaptive Hidden Markov Models, submitted to CVPR 2003.

• Xiaoming Liu and Tsuhan Chen, Shot Boundary Detection Using Temporal Statistics Modeling, Accepted for publication in the Proceeding of the IEEE International Conference on Acoustics, Speech, and Signal Processing 2002.

• Xiaoming Liu, Tsuhan Chen and B.V.K. Vijaya Kumar, On Modeling Variations For Face Authentication, Accepted for publication in the Proceeding of the International Conference on Automatic Face and Gesture Recognition 2002.

• Xiaoming Liu, Yueting Zhuang and Yunhe Pan. Video Based Human Animation Technique. In: Proceeding of The 7th ACM International Multimedia Conference (Multimedia 99). Orlando, Florida, USA. October 30-November 5, 1999.

• Xiaoming Liu, Yueting Zhuang and Yunhe Pan. A New Approach to Retrieve Video by Example Video Clip. In: Proceeding of The 7th ACM International Multimedia Conference (Multimedia 99). Poster Session. Orlando, Florida, USA. October 30-November 5, 1999.

• Yueting Zhuang, Xiaoming Liu and Yunhe Pan. Video Motion Capture Using Feature Tracking and Skeleton Reconstruction. In: Proceeding of IEEE International Conference on Image Processing 1999. Kobe, Japan. October 25-28, 1999.

• Xiaoming Liu, Yueting Zhuang, Yi Wu and Yunhe Pan. Video Based Human Motion Capture. In: Proceeding of IEEE Signal Processing Society 1999 Workshop on Multimedia Signal Processing. Copenhagen, Denmark. September 13-15,1999.

• Yueting Zhuang, Xiaoming Liu and Yunhe Pan. Apply Semantic Template to Support Content-based Image Retrieval. In: Proceeding of IS&T and SPIE Storage and Retrieval for Media Databases 2000. San Jose, California, USA. 23-28 January,2000.

          Technical Reports: 

• X. Liu, T. Chen and S. M. Thornton, "Eigenspace Updating for Non-Stationary Process and Its Application to Face Recognition", Carnegie Mellon Technical Report: AMP02-01, September 2002.

• Xiaoming Liu, Tsuhan Chen and B. V. K. Vijaya Kumar,  Robust Face Authentication for Multiple Subjects Using Eigenflow. Carnegie Mellon Technical Report, AMP 01-05. 

Top of this page