PERBANDINGAN DTCWT DAN NM F PADA FACE RECOGNITION MENGGUNAKAN EUCLIDEAN DISTANCE

Dual tree complex wavelet transform (DTCWT) iswidely used for representation of face imagefeatures. DTCWT is more frequently used than Gabor or Discrete Wavelet Transform (DWT) because it providesgood directional selectivity in six different directions. Meanwhile, non-negative matrix factorization (NMF) isalso frequently used since it can reduce high dimensional feature into smaller onewithout losing importantfeatures. This research focused on comparison between DTCWT and NMF as feature extraction and EuclideanDistance for classification. This research used ORLFaces database. Experimentalresult showed that NMFprovided better results than DTCWT did. NMF reached 92% of accuracy and DTCWT reached 78% ofaccuracy