Review on Writer Identification and Verification Methods

Biometric authentication requires identifying patterns based on unique behavioral or physiological characteristics. The most commonly considered physiological biometric systems are fingerprint scanning and facial recognition. Behavioral biometric systems are signature, handwriting, and voice recognition. Verifying physical features occurs efficiently through comparison, whereas recognizing behavioral components presents a challenge. Among the different biometric modalities, writer identification and verification have garnered considerable interest because of their potential to improve accuracy and reliability of authentication systems. Writer verification forms the legal basis for documents such as wills, contracts, and suicide notes. Handwriting verification tasks require years of training, and handwriting analysis is time-consuming. Therefore, researchers have tried to build accurate and efficient writer identification and verification systems. Researchers over the decades have explored various approaches for writer identification and verification tasks, including traditional machine learning methodologies and, more recently, deep learning methods. This review paper presents an informative summary of the databases, preprocessing methods, feature extraction techniques, and classifiers employed in identifying and verifying writers. While writer authentication has shown promising results, further research is needed to address several challenges in this task.