Results in which a fingerprint image has been processed to detect a skeleton image (top right), then associations between ridges and minutiae have been detected (left). The same ridges have been color-coded (bottom right) to aid in visualization.
Lynn Abbott and Michael Hsiao are on an interdisciplinary team devoted to quality assessment of fingerprint images. Automated fingerprint identification systems typically depend on detecting the locations of “minutiae,” which are the endpoints and bifurcations of friction ridges on a person’s skin. Human examiners, in contrast, usually rely on properties in addition to minutiae locations, such as the shape of ridges themselves. The team is developing a novel approach that involves “extended feature sets,” combining minutia and friction ridge properties with other data.