Performance comparison of DCT, FFT, WHT & Kekre's transform for on-line signature recognition

Dynamic signature is an important behavior based biometric. Dynamic features of human signature are available in case of on line signatures. Spatial Co-ordinates, pressure, azimuth, altitude variation w. r. t. time is analyzed in this paper. The signature feature vector is extracted from the captured feature points using transforms such as DCT, FFT, WHT & kekre's Transform. Derived features such as Velocity, Acceleration, velocity & Acceleration angle as well as Row & Column mean of pressure is used for analysis. Finally the performance is compared for above mentioned variations.