Handwritten Signature Verification using Deep Learning

Handwritten Signature Verification using Deep Learning

Ankush Anil Das, Arthikesh Himmat Patil, Vaibhav Muktyarsing Patil, Rajashree Rajendra Shinde, Prof. Puspendu Biswas

Computer Engineering
Sanghavi College of Engineering
Nashik, Maharashtra, India

ABSTRACT

Even though most areas, including land records, agreements between parties, legal certificates, identification cards, etc., are moving toward digital documents with digital signatures for authentication, uses only a signature written by hand. Verifying signatures is crucial because a false signature would have a significant impact on the actual owner. Therefore, it is essential to recognize genuine signatures in order to avoid such frauds. In this work, the deep learning technique is used to recognize the signature because it produces the highest accuracy and does not require excessive preprocessing. Image processing, classification, and segmentation are the most common applications for a deep learning model that is based on a convolutional neural network (CNN). The CNN algorithm learns more than KNN, SVM, and other algorithms. This work makes use of CNN to improve classification.

Keywords: Signature, Convolutional Neural Networks (CNN), Support Vector Machine (SVM), K-Nearest Neighbors (KNN).