Signature pattern recognition system (SPR system)

• Main Authors: Aparte, Alfred Angelo D., Cuento, Carlo V., Lopez, Paul Roderick B., Papa, Mary Ann V.
• Format: text
• Language: English
• Published: Animo Repository 1993
• Subjects: Pattern recognition systems; Signatures (Writing); Computer systems; Penmanship; Computer vision; x4 Handwriting
• Online Access: https://animorepository.dlsu.edu.ph/etd_bachelors/9514
• Institution: De La Salle University

The Signature Pattern Recognition (SPR) Software is a system that determines percentages of similarity in compared signatures, and renders it valid or invalid based on a predetermined decision result. Its main functions are: ADD SIGNATURE, DELETE SIGNATURE, and COMPARE SIGNATURES. The first two functions are used to maintain the pool of signatures stored on disk. The third capability concerns the usage of pattern recognition algorithms to deem whether two signatures are valid or invalid. In the comparison of signatures, qualities and distinctive characteristics of signatures were used as basis for the detection of similarity. Such were: the overall form, the baseline pattern, and the spacing between letters and words. Each method tolerates errors achieved during the signing and scanning process. All were chosen because these characteristics of signatures recognizes the uniqueness of each signature and tolerates errors which are based on the assumptions that no two signatures are exactly the same and a signature scanned twice will never yield the same image due to the human factor and the input device used. A survey was used to predict the behavior of the methods individually. The algorithms and theorems used were built upon the characteristics of signatures, thus the SPR system can determine similarity in form and certain signature attributes with some tolerance and flexibility. The decision results were based on the survey which determined a pattern of percentages on signatures made by the same person and those forged, to guide the decision result on similarity with authenticity.