摘要
针对现有掌纹分割算法中存在丢失部分有效信息的问题,提出了一种新的五点掌纹分割算法。首先,用基于仿射变换快速方法找到手掌的最大内切圆;然后,又用坐标变换和角度标定的方法找到手掌上的五个特征点;最后,通过对这五个特征点的处理实现手掌图像的分割定位。仿真结果表明,该算法准确、有效,且所得掌纹有效面积明显大于其他分割方法。
This paper proposed a novel five-point segmentation algorithm, considering the problem which had lost some effective information in existing palmprint segmentation algorithms. At first, the maximum inscribed circle should be found in palm by a fast method based on affine transformation. And then the five feature points would be gotten by the methods of coordinate transformation and angular calibration. At last palmprint segmentation could be realized by processing the five points. The results show that the algorithm is accurate and effective. And the effective area of palmprint is obviously bigger than those in other methods.
出处
《计算机应用研究》
CSCD
北大核心
2010年第6期2370-2372,2376,共4页
Application Research of Computers
基金
山西省自然科学基金资助项目(2008011030)
关键词
掌纹分割
五点分割法
角度标定
仿射变换
掌纹识别
最大内切圆
palmprint segmentation
five-point segmentation
angular calibration
affine transformation
palmprint recognition
maximum inscribed circle
