Jia Li

Jia Li
Assistant Professor of Statistics
Ph.D, Electrical Engineering, Stanford University, 1999

Summary of research interests

Jia Li's research interest includes statistical modeling for image processing, classification, and vector quantization.

Hidden Markov models have been used widely in many fields including speech recognition and genomic sequence analysis. However, those hidden Markov models are one dimensional and hence have limitations in processing information in higher dimensions. Dr. Li has constructed a two dimensional multiresolution hidden Markov model so that various image processing techniques may be improved by exploiting context information embedded efficiently in the model.

Dr. Li has explored applying vector quantization techniques to statistical classification. Although the popularity of vector quantization in technology arises primarily from data compression, it is also a natural statistical learning method. The well-known k-means clustering algorithm and the Lloyd VQ algorithm used in data compression are essentially the same. Studies have shown that algorithms aimed at reducing simultaneously the probability of misclassification and quantization error often achieve higher classification accuracy than systems designed solely for classification.

The current research of Dr. Li concerns the classification of high dimensional sparse data with applications in data mining and content-based image retrieval.

Biosketch

Jia Li was born in Hunan, China, in 1974. She received the B.S. degree in Electrical Engineering from Xi'an JiaoTong University, China, in 1993, the M.S. degree in Electrical Engineering in 1995, the M.S. degree in Statistics in 1998, and the Ph.D degree in Electrical Engineering in 1999, all from Stanford University.

She worked as Research Assistant on image classification and compression in the Electrical Engineering Department, and Research Associate on content-based image database retrieval in the Computer Science Department at Stanford University. She also worked at the Xerox Palo Alto Research Center on image processing. She is currently Assistant Professor of Statistics at Penn State University.

Representative publications

Jia Li and Robert M. Gray, Image Segmentation and Compression Using Hidden Markov Models, Kluwer Academic Publishers, 2000.

Jia Li, Robert M. Gray, and Richard A. Olshen, "Multiresolution Image Classification by Hierarchical Modeling with Two Dimensional Hidden Markov Models," IEEE Transactions on Information Theory, August 2000.

Jia Li and Robert M. Gray, "Context-based Multiscale Classification of Document Images Using Wavelet Coefficient Distributions," IEEE Transactions on Image Processing, September 2000.

Jia Li, Amir Najmi, and Robert M. Gray, "Image Classification by a Two Dimensional Hidden Markov Model," IEEE Transactions on Signal Processing," 48(2):517-33, February 2000.

Jia Li, Navin Chaddha, and Robert M. Gray, "Asymptotic Performance of Vector Quantizers with a Perceptual Distortion Measure," IEEE Transactions on Information Theory, 45(4):1082-91, May 1999.

Jia Li, James Z. Wang, and Gio Wiederhold, "Integrated Region Matching for Image Retrieval," Proc. ACM Multimedia, Los Angeles, Oct. 2000.