Fingerprint Feature Extraction Using Midpoint ridge Contour method and Neural Network

Automatic minutiae detection is an extremely critical process especially in low quality fingerprints. Most automatic systems for fingerprint comparison are based on minutiae matching. Minutiae are essentially terminations and bifurcations of the ridge lines that constitute a fingerprint pattern. A minutia extraction approach has been presented using the midpoint ridge contours. Our method do not requires thinning of fingerprint image but it can work with actual size ridges as it has been acquired. On the fingerprint image segmentation, normalization and contrast enhancement operations are required to be perform for clear identification of minutiae points. Color coding scheme has been used so that each ridge line should be scanned only once. Feature point locations that are being identified by midpoint ridge contours method are stored and passed to multi layer Perceptron trained with backpropagation algorithm. The result achieved is compared with those obtained through a method based on image thinning. The method proposed takes less time and do not detect any false minutiae.