Fuzzy feature selection for fingerprint identification

Fingerprint verification systems are expensive and complex, requiring sensing facilities, preprocessing algorithms for image quality enhancement and procedures for ridge and minutiae detection which can be used for classification, verification and recognition. To-date, the main published approaches to fingerprint recognition break down the process of ridge detection into smoothing or early pre-processing, edge detection, thresholding, binarization and subsequently thinning. This whole procedure can be very computationally expensive and hence requires more expensive hardware to meet the response-time requirements. The approach presented in this paper is based on fuzzy logic techniques. This has the advantage of being simple and less expensive.