Abstract— Recognizing hand-written signature is important due to several legal issues. Manual verification is often difficult when very much ‘similar-looking’ but forged signature is produced. In this work an effort has been made to automate such kind of signature verification process offline, using Adaptive Resonance Theory type-1 (ART-1). It is implemented using both serial and parallel processing, the performance of which are then compared. The said network has been trained with the original signature and tested with two forged signatures. The grade of similarity has been computed by introducing the term ‘Similarity Index’ (SI). Performance analysis reveals that after a careful tuning of vigilance parameter (ρ), both serial and parallel processing are able to learn the exemplary patterns with 100% accuracy. While testing, it is noted that parallel processing performs better than the serial processing in terms of speed as well as identifying the forged signatures by computing the mismatch.

Index Terms—-hand-written signature, automatic verification, ART-1, similarity index, forged signatures

Cite: Tirtharaj Dash, Tanistha Nayak, and Subhagata Chattopadhyay, "Offline Verification of Hand Written Signature using Adaptive Resonance Theory Net (Type-1)," International Journal of Signal Processing Systems, Vol. 1, No. 1, pp. 17-22,  June 2013 doi: 10.12720/ijsps.1.1.17-22