A Comparison ofPrincipalComponentAnalysis and GeneralizedHebbian Algorithm forImage Compression and FaceRecognition

Inthis paperweperform imagecompression andface recognition usingPrincipal Component Analysis (PCA)andthe Generalized Hebbian Algorithm (GHA)whichisoneofthePCA techniques involving neural network. Byimplementing thePCA andGHA algorithms forimagecompression wefoundthatPCA gives better compression ratio totheimagethanGHA andasfor facerecognition wefoundthatGHA gives morerecognition rate thanPCA. E(X) = 0 (1) R E RWm"isthecorrelation matrix oftherandomvector X, defined astheexpectation oftheouter product ofthevector X withitself asshown: