Feature Level Fusion using Multi-wavelet Based Iris Feature Extraction

Abstract

A new approach for the iris recognition based on feature level fusion using multi-wavelet transform is presented in this paper. It specifically uses the combined wavelet transform with multi-wavelet on the unique features obtained from the grey level iris images. It is composed of iris image acquisition, preprocessing, feature extraction and classifier design for matching process. The algorithm for iris feature extraction is based on texture analysis by using combination of wavelets and multi-wavelets transform. Multi-wavelet is extremely effective to analyze mutational and singular signals. It selects spatial directions and the energy is basically concentrated in low frequency section. Compared with existing methods, our method extracts 2-dementional information of iris which is scale, translation and rotation invariant. The fused iris image with combination of wavelets and multi-wavelets provide better accuracy and iris recognition rate.