Handwritten Signature Verification using CNN
Rutuja Nitin Lanjewar, Radha Wasudeo Wande, Swapnil K. Gundewar · 2024
A signature is a special kind of a sign made or given by a person on an instrument or document to indicate that he or she fully understands everything written on the document, agrees with them or being bound to fulfill the obligations stated on the document. It is used to verify the genuineness of the writing, to indicate origin and to constrain the signer to the provisions thus provided in the writing. Thus, the field of signature verification is not only valuable for the sphere of commercial banking, used for tools against frauds and approval of transactions, but also for a lot of other fields. Employment of this method is widely seen in legal, financial and governmental sectors to avoid forgery of numerous contracts agreements, legal instruments and other related documents. This paper concentrates on the offline handwritten signature verification using CNN, an advanced deep learning architecture which is famous for its capability to work on image data. CNNs are more appropriate for this task concerning the following characteristics: they are capable to draw out significant features from the province of the images of signatures, and the features like the strokes, texture, and general structure of signatures and so on with the high classification accuracy of genuine and forged signatures. This approach makes the signature verification more efficient and accurate and provides a better and automated solution for many applications for which signature authentication is most essential.