Implementation of Hand Written based Signature Verification Technology using Deep Learning Approach

Abstract

There have been several iterations of biometric techniques released for use in personal identification. Vision-based technologies include facial recognition, fingerprint scanning, iris scanning, and retina scanning. The most well-known non-visual methods are voice recognition and signature verification. As long as signatures remain an important element of financial, economic, and legal procedures, truly secure authentication is becoming more important. Signatures from authorized individuals continue to be the most trusted form of verification and are sometimes referred to as “seals of approval.” This research describes a method that begins with the processing of an image, continues with the extraction of geometric characteristics, and ends with the training and verification of a neural network. The authenticity of a test signature may be determined by feeding its extracted features into a neural network that has been trained as part of the verification process.