Candidate thesis offers unique improvements to handwritten signature recognition

With digitization and biometric ID systems on a rapid rise, recognition software becomes more and more important for our everyday lives.

Senior Lecturer Ellina Anisimova (Department of Mathematics and Applied Informatics, Yelabuga Institute) successfully submitted her PhD thesis in November 2020; the title was ‘Recognition of dynamic handwritten signatures based on methods of the fuzzy set theory.’

Anisimova explains the gist of her research, “A person enters a handwritten signature on a tablet PC. Various characteristics of the signature are then analyzed – dot coordinates, pressure strength, azimuth, angle, and others. Recognition of handwritten signatures has to deal with various peculiarities, such as instability, meaning that a person’s signature varies from one writing to another, and presence of high-quality forgeries. That’s why the fuzzy set theory was proposed to study dynamic signatures.”

She studied the effectiveness of her algorithm with MCYT_Signature_100, a database of signature specimens. The approach proved to have 0.36% of equal error rate, which is better than existing methods of recognition.

Anisimova started to research the topic in her master course at Kazan Federal University. Later, she enrolled in a PhD program at Kazan National Research Technical University.

“Signature recognition is extremely important for hybrid document turnover systems, electronic government, and banking,” adds the young scientist’s research supervisor, Professor Igor Anikin. “She not only put forth mathematical apparatus and algorithms to improve recognition in real-life situations, but also created a ready-for-use software tool for information systems.”

Source text: Yelabuga Institute

Translation: Yury Nurmeev