Image quality quantification for fingerprints using quality-impairment assessment

A quality impairment assessment along with a quality score would enable Automatic Fingerprint Identification Systems (AFIS) to make appropriate decisions to a) reject the fingerprint and recapture another sample, b) use other fingers or biometric features for recognition, c) use image enhancement techniques. Our approach provides a quality score in addition to a quality impairment assessment into dry, wet or small contact area fingerprints, using which the fingerprint could be rejected to re-capture another sample after wiping the finger/using additional pressure. A manual labeling of dry, wet and normal fingerprint regions in the FVC2002 DB1 database is used to create classifiers for the quality impairment assessment. A block based quality impairment classification approach is used to compute an overall image quality score. The block classification into dry, wet or normal blocks has 96.07% accuracy. The overall quality score is between -1(poor quality) and 1(excellent quality), which is found to be satisfactory through a visual inspection.