Chun-Jen Tsai ()
¤u¤TÀ]718«Ç,PH: 03-5731628


Current Research Area:
  • Multimedia SoC design
    Directions (sponsors: NSC, Novatek):
    - AVC (H.264/MPEG-4 Part 10) Codec SoC Design.
    - Application-Specific Instruction Set Processor for multi-standard video codecs.
    - Java processor and JRE design for embedded systems.
  • Software/hardware co-design for embedded multimedia systems
    Directions (sponsor: NSC):
    - SW/HW Co-design of Digital TV Terminal using Platform-based FPGA.
    - MHP middleware design for Digital TV.
  • Advanced video codec design
    Directions:
    - Distributed video coding theory development.
    - Psychovisual model-based video encoder design.
    - Wavelet-based video codec investigation.
  • Scalable distributed multimedia systems
    Directions (sponsor: NSC):
    - RD-optimzed scalable video streaming systems.
    - Streaming system protocol and architecture design.
    - Distributed video coding for peer-to-peer applications.

Old Research Area:
  • Surface reconstruction using passive sensors (CCD cameras).
  • Depth-from-Defocus.
  • Sensor integration for surface reconstruction.
  • Target tracking.
  • Disparity field estimation.
  • Total least-squares estimation.
  • Optical flow estimation in multi-channel video sequences.
  • Synthetic view generation with or w/o object models.

I haven't get the chance to dcument my current research here yet. However, as a start, the following old pages are some interesting works I've done before.

Asymptotic Bayesian Surface Reconstruction.
Disparity Estimation.
Multi-view Image Generation (follow this link to see a stereo image).
Model-based Synthetic View Generation for Video Conferencing.

Last update: Nov/30/2004

cjtsai@csie.nctu.edu.tw