中文摘要：

块状纹理是虹膜表面上一种重要的特征纹理。对于可见光虹膜图像而言,由于受眼睑、光斑及光照不均匀等因素影响,如何准确地提取出该特征纹理仍然是目前一个亟待解决的难题。提出一种基于分枝定界（BAB）策略和支持向量机（SVM）的可见光虹膜块状纹理检测方法。首先使用线性模板来提取虹膜上所有目标纹理;然后,采用BAB策略定位所有目标纹理并计算每个目标纹理的特征向量;最后,使用SVM实现对块状纹理的检测。通过实验证明,该方法对于虹膜块状纹理的检出率和检测时间均优于同类方法。

英文摘要：

Lump-like texture is an essential feature texture of iris. Due to the influences of eyelid,light spot and uneven illumination,to detect lump-like texture from visible iris image is still a challenging topic. A method based on BAB（ branch and bound） strategy and SVM（ support vector machine） algorithm is proposed to detect lump-like texture in this work. First,linear templates is applied to abstract all target textures in the iris. Then,BAB strategy is utilized to obtain the feature vector for each target texture that can describe the texture feature. Finally,the lump-like textures are detected by using SVM algorithm. Experimental results show that the proposed method is superior to other methods both in detection rateand operating time.