HYBRID APPROACHES COMBINING SOFT COMPUTING TECHNIQUES FOR HANDWRITTEN SIGNATURE

Handwritten signature verification is a critical problem in biometric authentication systems due to high intrapersonal variability and the presence of skilled forgeries.Traditional machine learning approaches such as Support Vector Machines (SVM) and Hidden Markov Models (HMM) often fail to generalize effectively under these conditions.Although Convolutional Neural Networks (CNNs) have demonstrated strong feature extraction capabilities, their rigid decision boundaries and lack of interpretability limit their robustness in uncertain scenarios.To address these challenges, this paper proposes a hybrid CNN-Fuzzy Logic framework for offline handwritten signature verification.The CNN component is employed to automatically learn discriminative deep features from signature images, while a Fuzzy Inference System (FIS) is used for decisionmaking to handle ambiguity and uncertainty in classification.The proposed hybrid approach improves adaptability to intra-personal variations and enhances detection of skilled forgeries.Experimental evaluation on the CEDAR signature dataset demonstrates that the proposed model outperforms several state-of-the-art methods in terms of accuracy, precision, recall, and F1-score, while also providing improved interpretability and robustness.