Fingerprint Segmentation: Optimization of a Filtering Technique with Gabor Filter

An Automatic Fingerprint Recognition system is dependent in many ways on the segmentation of the input fingerprint images because it improves the images so that features can be efficiently extracted from these images by the system. A noisy image tends to give a lot of false features to the feature extraction algorithm, hence segmentation is needed. Gabor Filters have proved their advantages in many applications of segmentation and enhancement. In this paper the researcher has tried to improve an existing algorithm for the segmentation of low quality fingerprint images by improving the filtering technique used by it. The reason for choosing Gabor filter being its property of optimal joint resolution in spatial and frequency domain. The optimized algorithm performs better with the changed filtering technique using Gabor filter. The algorithm works in three phases. The pre-processing phase makes use of Gabor Filter. Local features are used for the purpose of segmentation. In the final stage, a modified version of Otsu thresholding is applied in combination with fill the algorithm.