Baseline detection using Particle Swarm Optimisation

Signature is a popular method of seeking approval and authentication between various parties in many transaction applications. Signature pattern recognition is done by processing a set of data that consists of (x, y) coordinates, representing online signature. Particle Swarm Optimisation technique is used to find and analyse the baseline feature that exists within the signature. Signatures were taken from twenty randomly selected individuals with different backgrounds. We classified the baseline to three values, which are straight baseline, ascending baseline and descending baseline. The generated result of baseline feature was then compared to the truth value of baseline feature and our method achieved 90% accuracy. Thus, the proposed algorithm was found to be acceptable in representing the extraction of signature features based on baseline feature.