研究主題
1. Image/Video Compression and transmission :
- Image/video processing
- H.264/HEVC video encoder/decoder
- 3D AVC, 3D HEVC for 3D video
- 2D/3D video quality assessment (VQA)
2. 3D video and 3DTV-related :
- Stereo matching/depth estimation for two-view to multi-view conversion
- Automatic 2D-to-3D stereo video conversion
- Semi-automatic 2D-to-3D stereo video conversion
- Stereo and multi-view video coding
- Depth video coding
- DIBR (Depth-image-based rendering) technique for view synthesis
- Light field image processing
- Free viewpoint TV (FTV)
3. Video surveillance system :
- Machine learning/Deep learning approach for anomaly event detection
- Deep learning for human’s pose/action recognition
4. Medical Processing :
- Endoscopic (內視鏡) image/video analysis for MIS (Minimally Invasive Surgery, 微創手術)
- HDR (high dynamic range) imaging for endoscopic imaging
- 2D-to-3D video conversion for endoscopic video
- 3D pose estimation for instruments in MIS
5. ADAS (Advanced Driver Assistance System) applications :
- Stereo vision on the car
- Distance estimation of front vehicle and ACC (Adaptive cruise control) system
6. Intelligent Robot Vision :
- Deep learning for 6-axis industrial robot pick-and-place, random bin picking
- Deep learning for grasp position optimization
- Deep learning for object recognition and picking for humanoid/service robots
- RGB/RGBD sensing for industrial/service robots
7. Deep learning applications :
- Deep learning for 3D human skeleton extraction
- Deep learning for skeleton-based human action recognition/prediction
- Deep learning for 6DoF object pose estimation
- Deep learning for head pose estimation
- Deep learning for Wind/Solar power forecasting
- Deep learning for medicine Drug-Drug Interaction (DDI) and Adversary Reaction (ADR) prediction
8. Others :
- Algorithm/hardware design of infrared image tracker
- Multispectral/Hyperspectral image processing and recognition for ATR system
- VR/AR system in education/learning/exercising