Palmprint based Pattern Recognition Using Fast ICA

The pattern of Palmprint has high dissimilarity because of its numerous determination features like ridges, principle lines, datum points, and wrinkles and so on. In this paper, a blend of Haar Wavelet, DCT and Fast ICA have been utilized to recognize the palmprint pattern using fast ICA. The images of Palm were initially decomposed by the Haar Wavelet and frequency band, independent of different image resolutions chosen for dimensionality decrease. Next on LL, band DCT was use. Lastly, Fast ICA use to discover the independent components of the DCT features. To find matching image, Euclidean Distance (ED) has been used. The procedure is validated on Polytechnic- University (Poly U) storage as well as the results show the efficiency and accuracy of the proposed method.