Comparative Study of Image Matching Techniques for Secure Transactions in the Banking Sector

摘要

In the modern world all the banking process, become computerized for easy access to data. However, computerized data needs security for monetary transactions in banking. In banking, the security of account data is facing problems due to hacking or forgery of signatures. Saving accounts, partnership accounts are frequently accessed. In this paper, image matching techniques are analyzed and their performance is compared. This matching technique enables the account holders to perform transactions with each other's knowledge. Joint account holders have to provide their images individually at the time of the transaction. Images stored in the database and the ones provided are matched to approve the transaction. In case of a disability of any of the partners, it permits the other partners to authenticate themselves using the image of the deceased partner. By matching the images of all the account holders with images in the bank database, the transactions are approved. The transactions are denied when any one of the images mismatch. The image matching techniques Blob detection, Template matching, and surf extraction are implemented and their performance was compared. The performance of the Surf extraction technique was found to be best compared to the other two techniques.