The Common Vector Approach (CVA) has been introduced  especially for speaker independent speech recognition not long ago. The CVA is a subspace method based on calculation of the common vector for each class and the use of this vector in the recognition of classes. The common vector is a unique vector which represents the common properties of each class. CVA gave satisfactory results for the insufficient

data case (n: vector dimension ≥ m: number of vectors in each class) in the isolated-word recognition, speaker recognition and fault detection of motors. CVA was also applied

to the isolated word recognition for the sufficient case (n < m) and again satisfactory results were obtained.

̃ PCA

̃ FLD

̃ CV

̃ SVM

̃ PCA & SVM
Common Vector Approach

SSIP 2008 16th summer school on Image processing July 7th -16th 2008 Vienna

Gender Face Classification