Guest lecture for EE 148 (EE/CNS 150): Machine Learning for Computer Vision
by H.-T. Lin, April 2005.
Boosting is a popular method for Machine Learning, and has many applications in Computer Vision. In this guest lecture to the EE148 (EE/CNS 150) class, I would introduce the concept of boosting as intuitive as I could. The presentation shall start with the example of "Apple Recognition Problem", which is meant to be understandable even to elementary school students. Then, the audience should be able to get the essence of the famous Adaptive Boosting (AdaBoost) algorithm.
The second part of this presentation is from the paper "Sharing Features: efficient boosting procedures for multiclass object detection" by Torralba et al. in CVPR 2004. I would show their results through the figures, and discuss from a machine learning perspective.
Feel free to contact me: "htlin" at "csie.ntu.edu.tw"