Principal component analysis (PCA): PCA for gait recognition is
based on the information theory approach. It extracted the
relevant information in a gait image and encoded as efficiently as
possible. It identifies the subspace of the image space spanned by
the training image data and decorrelates the pixel values.