Machine Learning Algorithms for Detecting Fake Signatures in Bank
Milind Talele, Saurabh Palve, Anshika Saxena · 2025
Fake signatures pose a big risk to the safety and trustworthiness of bank operations it leads to financial loss which decreases in customer confidence and trust. This paper aims to delve into the issue of fake signatures in the banking world, shedding light on how it’s done, how far-reaching its effects are, and the proposed model to present it. It provides details review on what fake signatures are and why they matter in the context of bank fraud. We then look at the different ways fraudsters copy or change signatures to make unauthorized transactions. Moreover, the research looks at the current state of how signatures are checked and points out weak spots that make them open to fake signature attacks. This study also looks at progress in digital technology and how it can be used to better identify fake signatures, including machine learning models and biometric verification methods. By assessing how effective these tech solutions will help and suggest a way to reduce the risks tied to fake signatures in the banking sector including rules and safety measures which will help protect from frauds in future. Signatures are used to check the identity of the customer and authorize in bank, where daily financial transactions are checked using handwritten signatures. However, the common use of signatures has led to a rise in fraudulent activities, including signature forgery and disguises, where people try to copy the other signature. It provides the impact and correlation between perceived size of a handwritten signature and its graphical complexity, and other features in spotting fake signatures, thereby improving the trustworthiness and safety of banking transactions by proposing different AI models like CNN, SNN, ResNet 50 and indicate that SNN model perform well with the experimental results.