The project titled “API-Based Artificial Intelligence Supported Online and Offline Signature Verification System” developed by Fatih Mehmet Gürbüz, Hasan Şenmemiş, Ahmet Bilge and Zeynep Akalın, who are studying in the Computer Engineering and Electrical-Electronics Engineering departments of Erzurum Technical University (ETU), has been awarded support within the scope of the TÜBİTAK 2209-A University Students Research Projects Support Program.

The project, carried out under the consultancy of Dr. Nursena Bayğın, a faculty member of the Computer Engineering Department of the Faculty of Engineering and Architecture at ETÜ, and Dr. Sefa Küçük, a faculty member of the Electrical and Electronics Engineering Department, aims to detect both digital and physical signatures with high accuracy.

The system developed as part of the project utilizes advanced image processing techniques such as grayscale, thresholding, edge detection, and contour analysis for offline signature verification. For online verification, dynamic attributes such as speed, pressure, direction, and timing are analyzed during the signature creation process. The deep learning model, "SignatureCNN," developed by the project team, can distinguish between genuine and forged signatures with a high success rate. It also enhances the clarity of signature images by removing background noise from stamps, logos, and text, which are common in official documents.

The project team emphasized that the system will provide a reliable solution, particularly in areas with a high risk of fraud, such as banking, finance, law, healthcare, and education. The project team stated that the developed technology will increase security and significantly reduce fraud attempts by accelerating both digital and physical signature verification processes.

Corporate Communications and Promotion Directorate 11.08.2025