Ear recognition using features inspired by visual cortex and support vector machine technique
Ear recognition using features inspired by visual cortex and support vector machine technique
复制标题
DOI：
10.1109/iccce.2008.4580660
复制
发表时间：
2008-05
期刊：
影响因子：
--
通讯作者：
M. Yaqubi;K. Faez;S. Motamed
中科院分区：
文献类型：
--
作者：
M. Yaqubi;K. Faez;S. Motamed
耳朵是一类相对稳定的生物特征，从儿童时期到老年早期都保持不变。在大多数情况下，已经在其他生物特征领域应用的技术，比如主成分分析（PCA），也被应用于耳朵。本征耳（Eigen - ears）仅在严格控制的条件下才能提供较高的识别率。实际上，即使是轻微的旋转也会导致系统性能显著下降，而在无人值守的系统中旋转情况非常频繁。HMAX是一种特征提取方法，该方法是由视觉皮层的定量模型所推动的。此外，支持向量机（SVM）是在许多不同任务中都展现出高泛化能力的分类器，包括物体识别问题。在本文中，我们将这两种技术结合用于鲁棒的耳朵验证问题。我们利用北京科技大学（USTB）数据库来测试我们的方法。使用HMAX模型和支持向量机（SVM）分类器（核函数 = 1）组合的实验结果在耳朵验证中获得了比使用HMAX模型和k - 近邻分类器更高的识别率。此外，还证明了该方法具有旋转和尺度不变性，并且在实验中发现，在HMAX模型中使用高斯滤波器相比于使用加博尔（Gabor）滤波器提高了耳朵识别的性能。
Ear is a new class of relatively stable biometric that is invariant from childhood to early old age. In most cases techniques already working in other biometric fields, such as PCA are applied to ear. Eigen-ears provide high recognition rate only in closely controlled conditions. Indeed, even a slight amount of rotation can cause a significant drop in system performance and in unattended systems rotations occur very frequently. HMAX is a feature extraction method and this method is motivated by a quantitative model of visual cortex. Also, SVMs are classifiers which have demonstrated high generalization capabilities in many different tasks, including the object recognition problem. In this paper we combine these two techniques for the robust Ear verification problem. The USTB database is exploited to test our approach. Experimental results using the combination HMAX model and support vector machine (SVM) classifier (with kernel=1), obtains higher recognition rate than those obtained with HMAX model and k-nearest neighbors classifier in ear verification. In addition to, demonstrated that this method is rotate-and scale-invariant, and also, in experiment, it was found that, using of Gaussian filter in HMAX model in compared to using of Gabor filter, increases performance of ear recognition.