MENU

Fun & Interesting

PCA 10: eigen-faces

Victor Lavrenko 115,910 lượt xem 11 years ago
Video Not Working? Fix It Now

Full lecture: http://bit.ly/PCA-alg
We can perform PCA on photographs of faces. First we unfold each bitmap into one big vector. We run PCA and find principal components (eigenvectors) which represent salient properties of faces. These eigenvectors can be folded back into a bitmap, which can be visualized and are called eigenfaces.

Comment