HONG KONG BAPTIST UNIVERSITY
FACULTY OF SCIENCE

Department of Computer Science Seminar
2014 Series

Manifold Learning for Multimedia Data Analysis

Dr. Yang Liu
Department of Computing
The Hong Kong Polytechnic University

Date: October 28, 2014 (Tuesday)
Time: 11:30 am - 12:30 pm
Venue: SCT909, Cha Chi Ming Science Tower, Ho Sin Hang Campus

Abstract
Multimedia data analysis plays an important role in many real-world applications, such as image classification, face identification, film editing, and music therapy. How to represent these high-dimensional data in a compact and meaningful way is one of the most challenging problems. In this presentation, I will present a series of novel machine learning methods under the manifold learning framework to discover the concise and effective representation of multimedia data. Four components are considered in my work: 1) how to extract the low-dimensional manifold structure of the data sequences; 2) how to preserve the naturally high-order structure of the data points; 3) how to measure the distance between different data points; and 4) how to consider the correlations between different labels during the learning procedure. We validate the effectiveness of proposed methods on standard datasets in various multimedia data analysis tasks such as video trajectory description, image classification, and music emotion analysis. By referencing the human perception procedure, the presented techniques achieve impressive performances.

Biography
Dr. Yang Liu is the Coordinator of Cognitive Computing Lab in the Department of Computing at The Hong Kong Polytechnic University. He received his Ph.D. degree from the same department in 2011. After that, he spent one year as a Postdoctoral Research Associate in the Department of Statistics at Yale University. Dr. Liu's research interests include pattern recognition and brain modeling, as well as their applications in computer vision and music therapy. As the first author, he has published more than twenty papers on reputable journals such as IEEE Trans. Neural Networks and Learning Systems, Pattern Recognition, and top conferences such as AAAI, SIGMM. His paper named “Bidirectional visible neighborhood preserving embedding” achieved the Best Student Paper in ACM ICIMCS 2009.

********* ALL INTERESTED ARE WELCOME ***********
(For enquiry, please contact Computer Science Department at 3411 2385)

http://www.comp.hkbu.edu.hk/v1/?page=seminars&id=307&lang=tc
Photos  Slides