Development of Chimp Optimization Algorithm and Adaptive Shuffle Attention Net for Automated Bank Cheque and Signature Verification

Abstract

Processing of cheques has evolved into congestion that can impede the financial system’s efficiency as a result of the banking industry’s growing digitalization. The verification procedure has historically been done manually, with a bank staff visually inspecting the cheque for fraud or other problems. However, this procedure can be costly, requires more time, and is prone to mistakes. Therefore, the automated check verification system developed, which uses computer vision and deep learning algorithms to extract related data from images of checks and evaluate their legitimacy, is presented as a solution to these problems. Thus, this research work involves acquiring the cheque image from standard databases, and it is followed by the denoising stage to remove any noise or unwanted artifacts. Preprocessing is done by Contrast- Limited Adaptive Histogram Equalization (CLAHE) and median filtering process to remove the unwanted noises from the images. Subsequently, OCR (optical character recognition) is applied to extract the numbers from the cheque. Once the numbers are extracted, they are fed into a handwritten recognition system, which is done by Adaptive Shuffle Attention Net (ASAN). After the handwritten characters are recognized, the deep features of the signature are extracted ASAN. Here, the updated random vector-based chimp optimization algorithm (URV-COA) is used for optimizing the parameters in ASAN. Finally, a similarity check is performed between the extracted signature and the signature on the file stored in the dataset. If the signatures are sufficiently similar, the cheque can be verified as authentic.