摘要
研究了高分辨率掌纹图像的分割方法问题,将均值、标准差作为标准,实现了由纹线组成的掌纹图像的前景分割。对于手指部分的干扰,将其分为孤立的手指部分和与手掌相连的手指部分分别进行处理,利用区域生长法提取最大连通域达到去除孤立手指部分的目的;利用轮廓跟踪法比较CD、PD距离,定位与手掌相连的手指部分的位置,从而实现将其去除的目标。实验结果表明,该方法很好地实现了高分辨率掌纹图像的分割,为后续的特征提取工作奠定了基础。
Palmprint image segmentation is of great significance in research as an important step in palmprmt recognition process. Here we are concerned with research on high-resolution palmprint image segmentation approaches. Also, we use mean and standard deviation as a standard implementation for foreground segmentation of grain-made palmprint image. Regarding passive impact from finger region, we deal with isolated finger region and palm-connected finger region separately. Isolated finger regions are removed with region growing algorithm which can find the maximum connected component. The two distances, CD and PD, which are calculated and compared by figure tracking algorithm, are used to locate the palm-connected finger regions and remove them. Through experiments, our approach has successfully segmented high-resolution palmprint images, and provided a good foundation for the subsequent palmprint feature extractions.
出处
《新乡学院学报》
2009年第4期45-48,共4页
Journal of Xinxiang University
关键词
生物特征识别
掌纹识别
掌纹图像分割
边缘跟踪
biometrics
palmprint recognition
segmentation of palmprint
counter tracing
