Iris feature extraction and matching based on Nonsubsampled Contourlet transform
Xiaoyu Huang · Jisuanji gongcheng yu sheji · 2009
A novel feature vector extraction method is provided that combine non-parameter filters with the iris image for region of in- terested (ROI) for obtain robustly the iris texture feature vector of high-quality. The framework and performances of the filter based on the nonsubsampled contourlet transform (NSCT) are analyzed, the design step and the design key point of the non-parametric NSCT filter are provided. The preprocessing technology of the iris pattern for region of interested (ROI) of extraction is presented using coordinate conversion, cut back, translation, and combination. Filtering the iris ROI with the NSCT filter, getting the iris pattern feature vector which with rich information and few interference. Experiment demonstrates this method is more effective.