Image segmentation based on scanned document and hand script counterfeit detection using neural network
Ravi Babu Devareddi, R. Shiva Shankar, K. VSSR Murthy, Ch. Raminaidu · AIP conference proceedings · 2022
Biometrics could be considered as couple of behavioral instances signature verification, keystroke activity and physiological activities such as iris, fingerprint. In the past, before utilization of computers, hand script signature is popular as biometric authentication and it has become a common identity for personal verification in all sectors and especially instant verification required to permit bank cheques. Among different sectors most of them uses offline verification. Taking this as an advantage the signatures are being forged for document authentication. Due to this, there is a need for automation to verification of offline signatures. In this work developed a model for signature verification using image segmentation technique for image processing using Convolutional neural networks. This model is utilized to classify the signature as genuine which is leads to the claimed personal or forgery which is created by someone else. For evaluation of the performance of the system, here considered a test set which consists of genuine signatures and forgeries. And hence the results are declared whether the considered signatures are genuine or forged signatures.