CAJ | 학술논문

针对指纹低质量指纹图像分割的特征,并总结了常见指纹分割算法存在的缺点,引出采用边缘检测的方法进行低质量指纹分割算法的设计研究.本文考虑由于单纯采用Sobel-Snake算子仅能起到平滑和抑制噪声的作用,并不能实现指纹图像和背景的分离,故此在Sobel-Snake算子检测之前采用二维最大熵的方法对图像进行二维最大熵技术,然后通过Sobel-Snake算子获取边缘图像,并根据给定阈值向量实现了低质量指纹分割算法.最后给出实验效果和分割图像的评价指标.
침대지문저질량지문도상분할적특정,병총결료상견지문분할산법존재적결점,인출채용변연검측적방법진행저질량지문분할산법적설계연구.본문고필유우단순채용Sobel-Snake산자부능기도평활화억제조성적작용,병불능실현지문도상화배경적분리,고차재Sobel-Snake산자검측지전채용이유최대적적방법대도상진행이유최대적기술,연후통과Sobel-Snake산자획취변연도상,병근거급정역치향량실현료저질량지문분할산법.최후급출실험효과화분할도상적평개지표.
For the characteristic fingerprint of low-quality fingerprint image segmentation, and summed up the shortcomings of the common fingerprint segmentation algorithm leads to the edge detection method for the study of low-quality fingerprint segmentation algorithm. In this paper, we consider the simple Sobel-Snake operator can only play the two-dimensional maximum entropy smoothing and noise suppression effect, and cannot achieve the separation of the fingerprint image and the background, and therefore before the Sobel-Snake operator to detect two-dimensional maximum entropy method of image technology, and then obtain the edge images by Sobel-Snake operator, and low-quality fingerprint segmentation algorithm to achieve a given threshold vector. Final y, the experimental results and evaluation of image segmentation.