Derin Öǧrenme Tabani İmza Tanima Uygulamasi

Nowadays, with the increase of biometric studies, the diversity of biometric data increases and new methods are used in evaluation methods. Traditional biometrics, such as face, fingerprints, handpieces, now leave their place to a variety of biometrics, which contain characteristic information about more people and include movement information. In this study, the performance of the deep learning method based on convolutional neural network (CNN) is demonstrated on a nonlinear signature recognition problem. In this non-real-time signature recognition application, it has been tried to reduce the process load and memory requirement by using deep learning method. Two data sets with different participant numbers were created in the study. The performance and reliability of the system are examined by various ratios of training and testing data on these data sets.