Touchless palmprint and fingerprint recognition

Biometric systems based on hand traits captured using touchless acquisition procedures are increasingly being used for the automatic recognition of individuals due to their favorable trade-off between accuracy and acceptability by users. Among hand traits, palmprint and fingerprints are the most studied modalities because they offer higher recognition accuracy than other hand-based traits such as finger texture, knuckle prints, or hand geometry. For capturing palmprints and fingerprints, touch-less and less-constrained acquisition procedures have the advantage of mitigating the problems caused by latent prints, dirty sensors, and skin distortions. However, touchless acquisition systems for palmprints and fingerprints face several challenges caused by the need to capture the hand while it is moving and under varying illumination conditions. Moreover, images captured using touchless acquisition procedures tend to exhibit complex backgrounds, nonuniform reflections, and perspective distortions. Recently, methods such as adaptive filtering, three-dimensional reconstruction, local texture descriptors, and deep learning have been proposed to compensate for the nonidealities of touchless acquisition procedures, thereby increasing the recognition accuracy while maintaining high usability. This chapter presents an overview of the various methods reported in the literature for touchless palmprint and fingerprint recognition, describing the corresponding acquisition methodologies and processing methods.