基于灰度差分和脊线检测算子的虹膜裂缝检测

Iris cracks detection based on gray level difference and ridge line detection operator

虹膜裂缝是用于虹膜识别的特定形状虹膜纹理之一。为了解决在复杂背景下提取裂缝特征的问题,在对虹膜裂缝特征进行分析的基础上,提出了一种虹膜裂缝检测方法。首先用灰度差分模板得到裂缝纹理候选区域,并根据裂缝的先验知识筛选裂缝候选像素,在原图上分割成小区域;其次,采用脊线检测算子在小区域范围内对各种纹理进行检测;最后在二值化图像上计算该区域能量密度,排除非裂缝纹理,实现裂缝纹理的检测。该方法在图库中人工标定的1 201处裂缝中检测正确检出率达到94.09%。实验结果表明,本文方法在不需要手动去掉卷缩轮内区域的情况下,能够在多种虹膜纹理并存的复杂背景下检测出虹膜裂缝纹理。

Iris crack is one of the specific shape iris textures that can be used in iris recognition.In order to solve the problem of how to extract the iris crack characteristics in complex background,an iris crack texture detection method is presented in this paper on the basis of analyzing the iris crack characteristics.Firstly,the gray level difference template is used to obtain the candidate crack texture region, then the candidate crack pixels are screened out according to the prior knowledge of the cracks.The original image is divided into small blocks.Then,the ridge line detection operator is used to detect various textures in the small blocks.Finally,the energy densities of the blocks on the binarized image are calculated;the non-crack textures are eliminated,and the crack textures are detected.The method was used to detect the 1 201 cracks that were artificially marked in the image gallery;and the detection accuracy is 94.09%.The experiment results show that without artificially removing the collarette region,the proposed method can detect the iris crack textures from the visible iris image with complex background and multiple texture coexisting.