Iris Recognition Based on Matching Pursuits
Yueh-Shiun Lee, Jinn Ho, Wen-Liang Hwang, Chung‐Lin Huang · 2008
We propose a novel dynamic programming matching pursuit (DPMP) algorithm for iris recognition. The method modifies the matching pursuit (MP) algorithm to select the most representative atoms for the iris recognition. In the experiments, we demonstrate that our system has (1) better performance for both personal identification and verification, and (2) a better ROC curve, (3) less computation than the conventional MP-based iris recognition.