Author Description

With their distinctiveness and stability over time, fingerprints continue to be the most widely used anatomical characteristic in systems that automatically recognize a person's identity. This fully updated third edition provides in-depth coverage of the state-of-the-art in fingerprint recognition readers, feature extraction, and matching algorithms and applications. Deep learning (resurgence beginning around 2012) has been a game changer for artificial intelligence and, in particular, computer vision and biometrics. Performance improvements (both recognition accuracy and speed) for most biometric modalities can be attributed to the use of deep neural networks along with availability of large training sets and powerful hardware. Fingerprint recognition has also been approached by deep learning, resulting in effective and efficient methods for automated recognition and for learning robust fixed-length representations. However, the tiny ridge details in fingerprints known as minutiae are still competitive with the powerful representations learned by huge neural networks trained on big data. Features & Benefits: Reflects the progress made in automated techniques for fingerprint recognition over the past five decades Reviews the evolution of sensing technology: from bulky optical devices to in-display readers in smartphones Dedicates an entire new chapter to latent fingerprint recognition, which is nowadays feasible in “lights-out” mode Introduces classical and learning-based techniques for local orientation extraction, enhancement, and minutiae detection Provides an updated review of presentation-attack-detection techniques and their performance evaluation Discusses the evolution of minutiae matching from rich local descriptors to Minutiae Cylinder Code Presents the development of feature-based matching: from FingerCode to handcrafted textural features to deep features Reviews f