A Stable Minutia Descriptor based on Gabor Wavelet and Linear Discriminant Analysis
This work addresses the need for accurate and efficient fingerprint recognition systems, particularly in challenging conditions like poor image quality, but it appears incremental as it builds on existing descriptor methods.
The authors tackled the problem of improving fingerprint recognition by developing a robust minutia descriptor using Gabor wavelet and linear discriminant analysis, which achieved high effectiveness in experiments on FVC2004 and FVC2006 databases.
The minutia descriptor which describes characteristics of minutia, plays a major role in fingerprint recognition. Typically, fingerprint recognition systems employ minutia descriptors to find potential correspondence between minutiae, and they use similarity between two minutia descriptors to calculate overall similarity between two fingerprint images. A good minutia descriptor can improve recognition accuracy of fingerprint recognition system and largely reduce comparing time. A good minutia descriptor should have high ability to distinguish between different minutiae and at the same time should be robust in difficult conditions including poor quality image and small size image. It also should be effective in computational cost of similarity among descriptors. In this paper, a robust minutia descriptor is constructed using Gabor wavelet and linear discriminant analysis. This minutia descriptor has high distinguishing ability, stability and simple comparing method. Experimental results on FVC2004 and FVC2006 databases show that the proposed minutia descriptor is very effective in fingerprint recognition.