在本論文中，我們提出一個指紋奇異點偵測系統，此系統能正確的找出中心點與三角洲所在的位置。本論文主要是結合zero-pole model 和Gabor transform 來偵測指紋中的奇異點。我們利用zero-pole model 的特性，可以有效偵測出指紋中奇異點所在的區塊，再針對這些候選區塊，利用Gabor transform 來偵測區塊中奇異點真正的像素位置。此系統將可以有效的偵測到奇異點，使得指紋比對系統能更精確地克服旋轉與平移的問題，並且加快比對效率。所以所提出的奇異點偵測系統，對往後的指紋比對作業奠定更穩固的基礎。
In this paper, we propose a system of singular point detection in fingerprints. The system can detect core and delta accurately. The paper combines zero-pole Model and Gabor transform to detect singular points. At first, we apply zero-pole model to find efficiently some candidate regions including singular points. Then we use Gabor transform to detect singular points in pixels. This system can overcome the problems from rotation and translation precisely and efficiently. So the proposed method will result in more stable work for fingerprint recognition.
