Abstract: In this paper, feature vector is extracted from the online signature based on Hybrid Wavelet Transform II and then they are given to Hidden Markov Model based classifier. The samples and various combinations of Feature vector are tested and the system is analyzed based on their performance. On the first 128 samples of the pressure parameter Hybrid Wavelet Transform II was applied and 1-16 samples of the output were used as feature vector for signature verification. The performance of the system was compared by using Hidden Markov Model(HMM) classifier. The Left to Right topology of HMM was used as classifier. Considering Left to Right model for 1 – 16 samples KEKRE 128 offers best performance of FRR ranging from 5% to 11% . Considering Left to Right model for1 – 16 samples DCT 128 offers best performance of EER 4%.

Keywords: HMM Hybrid Wavelet Transform Online Signature Verification.

Open chat