Handwriting Analysis for Bank Cheque Verification Using EfficientNet

Cheque analysis and verification is a challenging problem in financial technology that still has yet to have solutions that are competent and efficient, and most importantly, accurate at detecting fraudulent cheques. This paper proposes the approach with the help of a convolutional neural network for handwriting analysis for bank cheque verification. The development of an automated system that can verify the authenticity of bank cheques by comparing the handwriting on the cheque with the handwriting of the account holder on record as well as handwriting samples within the cheque itself. We extract handwriting features from the cheque and compare them with the handwriting samples from various parts of the cheque as well as those within the bank database to ascertain feature similarity and report anomalous samples. This method provides a promising approach for the automated verification of bank cheques enabling an improvement in efficiency while enhancing the security and reliability of the cheque verification pipeline employed by various financial institutions.