Current projects



 “Personalized Photograph Ranking and Selection System Considering Positive and Negative User Feedback”, ACM TOMCCAP (Transactions on Multimedia Computing, Communications and Applications), pp. 1-20, Vol. 10, Issue 4, Article Number 36, June 2014.

An aesthetic prediction model is learned from labeled photographs by using the proposed image features and RBF-ListNet learning algorithm. The experimental results show that the proposed framework outperforms in the ranking performance: a Kendall's tau value of 0.432 is significantly higher than those obtained by the features proposed in one of the state-of-the-art approaches (0.365) and by learning based on support vector regression (0.384). To realize personalization in ranking, three approaches are proposed: feature-based approach allows users to select photographs with specific rules; example-based approach takes the positive feedbacks from users to re-rank photograph; list-based approach takes both positive and negative feedbacks from users into consideration. User studies indicate that all three approaches are effective in both aesthetic and personalized ranking.


Automatic Chinese Food Identification and Quantity Estimation,  ACM Siggraph Asia 2012 Tech Brief

Android App (2014/6),  Cfood: Chinese Food Recognition”

(Chinese Food Menu Recognition Based on Contextual Thesaurus)

Users can take photographs of the menus at hand and be able to know more about the foods and order them by identifying the menu texts using the app. Furthermore, brief nutrition information about the dish is also available in the app.






Personalized Photo Ranking and Selection System   ACM Multimedia 2010



We have proposed a novel personalized ranking system for amateur photos. While the features used in our system are similar to previous works, new features such as texture, RGB color, portrait (through face detection), and black-and-white are included for individual preferences. We show that (1) The performance of our system in terms of precision-recall diagram and binary classification accuracy (93%) is close to the best results to date for both overall system and individual features. (2) Two personalized ranking user interfaces are provided: the feature-based and example-based. Both are effective in providing personalized preferences, and twice more people prefer example-based in our user study

Video Stabilization using Robust Feature Trajectories,  ICCV 2009 (oral presentation)


This paper proposes a new approach for video stabilization.


Facial Performance Synthesis using Deformation-Driven Polynomial Displacement Maps

ACM Transaction on Graphics (Siggraph Asia), 2008





Cubical Marching Squares 2005
Research on surface extraction from volume data.



Domain Connected Graph: The Essential Skeleton of a 3D shape
A new representation for the skeleton of an arbitrary shape is proposed. The nodes of the graph are named domain points (significant points inside the shape) which can represent the local properties of a shape. The domain points are classified into three categories: joint points, end points and connection points.

Domain Connected Graph


A Tool For Alignment of Molecules

A Tool for Alignment of Molecules
A novel tool is proposed to align two molecules (not just proteins) based on their 3D structural data, and the user can observe the result of alignment visually via the tool. Most existing tools are designed only for alignment of proteins. Here, a new tool is developed to address shared structural features between protein structures and tRNA structures, that is, molecular mimicry, although they are two very different types of molecules.


3D Model Retrieval
A 3D model retrieval system based on Multi-View LightField and Multi-resolution Reeb Graph

3D Model Retrieval


3D Protein Model Retrieval

3D Protein Model Retrieval  2004
A 3D Protein Model Retrieval System based on Multi-View LightField method.


Chinese Painting Rendering
An Interactive Haptic Chinese Painting System



Virtual Sculptor

Virtual Sculptor
A Feature Preserving Haptic Modeling System


Realistic 3D facial animation parameters
Accurate 3D facial motion trajectories are reconstructed from mirror-reflected multiview videos, and the estimated data are also applied to synthetic face for realistic facial animation.




MPEG4 Synthetic/Natural Hybrid Coding
Our focus is on head/face motion tracking using three feature points, as well as efficient 3D modeling of individuallized face with textures. Currently we are working on "Virtual Talk" and Model-Based Video Conferencing System.


Video VR
An authoring system is proposed to construct panoramic images of real-world scenes from video clips automatically. Instead of using special hardware such as fish-eye lens

Video VR


VR Talk

VR Talk
Web-enabled speech driven facial animation. In this project, a speech-driven talking head is developed as a plug-in for web browsers (plug-in for Navigator; ActiveX for IE). Since it is a model-based approach, the bitrate requirement of this lifelike animation is only 14Kbps.


Photo VR
A prism-like photo-realistic environment map rendering system.

Photo VR



Image Talk
synthetic faces can be used in virtual communication such as virtual meeting, virtual video e-mail etc.; the technique can also be a new friendly user interface in various applications


MPEG-4 Rendering
A rendering module for MPEG-4 system module.


  Previous projects


The SpaceWalker building walkthrough system
As part of our graphics research into virtual worlds, and also stimulated by the construction of a new building for the Department of Computer Science and Information Engineering at National Taiwan University, our goal is to construct a walkthrough system, the SpaceWalker, that enables a user to explore the computer model of the department building. --Dr. Yuong-Wei Lei


Hair Rendering
Hair has some features that make it hard to be created by the computer. These features include a large number of hairs and complex physical properties of hair. In this work, we aim to develop a procedure to reproduce the beauty of hair. --Tzong-Jer Yang

Hair Rendering


JAVA VRML applet

The VRML group works for the development of a VRML applet which is able to display VRML world in our Multimedia Shared Browser as well as in other commercial browsers. -- VRML group


TheJavaGL 3D Graphics Library for Java
JavaGL is a 3D graphics library for the JavaTM programming language with an API which is similar to that of OpenGLTM. This 3D graphics library can help ones to develop 3D graphics applications on the Internet/Intranet, and is developed by pure JavaTM programming language. Now, JavaGL is available for download and some examples can be run online. --Robin Bing-Yu Chen


Medical Image Processing

Tread-mill: a VR device of walk-through system
3D Trackers
3D Sound
Motion Prediction
Gesture Recognition
Motion Chair: a haptical device
Force Feedback Joystick: a haptical device
A Shift-Resisting Public Watermark System for Protecting Image Processing Software
Time Shifting Resistant Digital Audio Watermarking Using Template Matching

Copyright c 1996,1997, Communications and Multimedia Laboratory, All rights reserved.
Last Updated: 2006-03-30