本論文提出一個虹膜辨識系統的設計。具體而言，我們所研擬的辨識方法可區分為四大步驟，包括瞳孔中心定位、虹膜區域擷取、特徵萃取與編碼以及編碼比對。首先，運用影像處理技巧將影像中的瞳孔部分獨立出來，再用Canny邊緣偵測方法得到所有瞳孔邊緣的座標，據以訂出中心座標。接著以此中心座標為參考點，擷取出環狀的虹膜區塊影像。再藉由二維小波轉換萃取虹膜特徵並與一維log-Gabor 濾波器作旋積，接著利用相位量化的方式進行編碼。最後，比對決策部分則使用漢明距，由其訂出決策結果。我們使用CASIA(Institute of Automation ,Chinese Academy of Science)提供的虹膜資料庫圖片進行驗證，模擬結果證實所研擬的虹膜辨識方法有著極高的辨識率。 In this thesis, the design of an iris recognition system is presented. Specifically, the proposed method consists of four major steps, which includes pupil center positioning, iris data acquisition, feature extraction and encoding, and code matching. First, a frame containing the pupil is singled out by image processing technique. Then the coordinates of all the points on the edge of the pupil are computed using Canny edge detector. Accordingly, the position of the pupil center is determined. Next an annular region of the iris image is taken with reference to the pupil center. A two-dimensional wavelet transform is performed for feature extraction. After convolution with a two-dimensional log-Gabor filter, the data is encoded by phase quantization technique. Finally, the matching process is conducted using Hamming distance measure. Using the database provided by CASIA (Institute of Automation, Chinese Academy of Science), the simulation result demonstrates that the proposed iris recognition system exhibits high recognition rate.
