Classification via Dimensionality Reduction for Computer Vision Applications
- 2006-04-19 (Wed.), 10:30 AM
- 二樓交誼廳
- 劉 庭 祿
- 本院資訊所副研究員
Abstract
We investigate several computer vision problems using classification techniques based on dimensionality reduction. Our methods are derived from manifold-learning viewpoints, and have many practical advantages for addressing vision applications such as face/object recognition, real-time visual tracking, and image retrieval. The presentation will include the key ideas of our approaches as well as comprehensive examples to demonstrate the efficiencies of the proposed algorithms.
最後更新日期: