DeepSign – Intelligent Signature Verification and Fraud Detection

Signature verification plays a vital role in banking, legal documentation, and digital authentication systems. The increasing number of forgery cases has created a strong demand for automated and intelligent verification mechanisms. This paper presents DeepSign, a hybrid system that combines traditional machine learning and advanced deep learning techniques to accurately distinguish between genuine and forged signatures. Handcrafted features such as HOG, LBP, and GLCM are extracted and classified using SVM and Random Forest. Deep CNN architectures such as ResNet, VGG, and Inception are employed for automatic feature learning. Autoencoders are used for anomaly detection. The hybrid approach improves robustness, scalability, and detection accuracy.