However, with growing usage of fingerprint in e-commerce applications and advent of smart mobilephone smatching fingerprints across different resolutions is also gaining importance. Further, there are fingerprint sensors, such as Lumidigm Venus that utilize information from multiple spectrums for fingerprintcapture. Matching such images with the ones obtained from optical or capacitive sensors requires additional research
Similar to other data driven research areasadvancements in fingerprint recognition, especially in the academic community, are dependent on the availability of large databases. Some of the large publicly available fingerprint databases include card ink-print databases, live scan fingerprint databases, multisensor fingerprint databases, multi-resolution fingerprint
databases latent with corresponding full fingerprint databases, and other special databases. A comparative analysis of all the existing public fingerprint databases is provided in Table 1, which also provides a listing of the types of research challenges that can be addressed using each database. Existing databases primarily have two limitations:they generally only contain image variationscorresponding to a few challenges, and some challenges such as latent fingerprint recognition
and cross spectral matching have small databases associated with them.