A Statistical Approach for off - Line Signature Verification

This paper includes off line Signature Verification (SV) process with test results using the proposed algorithm Particle Swarm Optimization-Neural Network (PSO-NN) together with statistical analysis, Chi-square test. The verification process is performed in four steps. Signature images are scanned (data acquisition) and image processing is applied to make images suitable for extracting features (pre-processing). Each pre-processed image is then used to extract relevant geometric parameters (feature extraction) that can distinguish signatures of different volunteers. Finally, the proposed verification algorithm is tested on the database that includes 1350 skilled and genuine signatures taken from 25 volunteers. The Chi-square test is applied to see how the signature data fits with probability test function.