HANDWRITTEN SIGNATURE VERIFICATION USING DEEP LEARNING

HANDWRITTEN SIGNATURE VERIFICATION USING DEEP LEARNING

Prof. M. V. Korade 1, Neha Porje2, Rasika Borade3, Siddhika Mahajan4, Janhvi Patil5

1,2,3,4,5 Sandip Foundation’s Sandip Institute of Engineering and Management/

Computer Engineering/Savitribai Phule Pune University/India

---------------------------------------------------------------------***---------------------------------------------------------------------

Abstract - Signature is most generally used for verification of an individual or a private. Signature is considered as a mark for the identification of all the social and business functions. The term signature verification is of utmost importance because it may be victimized and might result in Brobdingnagian losses. The signature could be a behavioral biometric attribute that includes neuromotor characteristics of the signer (e.g., our brain and muscles among others outline the means we tend to sign) in addition as socio-cultural influence. Throughout centuries, the examination of signatures has been created by consultants United Nations agency confirm the credibleness of the sample supported rhetorical analysis. During this analysis we have used forensic options for identification of author like angle, categorical attributes, scalar measures, alignment to the baseline, length of strokes, slant of strokes, shape, punctuation, order, text loops, character spacing, etc. These factors that facilitate within the analysis of the verification of the signature primarily facilitate to find the author conjointly. For the analysis of signature an oversized dataset is needed for the experimentation purpose and for testing purpose. Thus, the datasets for the signatures are taken for more analysis.

Key Words:  Deep Learning, Signature, Verification, Analysis.