A vehicle driver iris detection algorithm
A vehicle driver iris detection algorithm
复制标题
一种车辆驾驶员虹膜检测算法
DOI：
--
复制
发表时间：
2009
期刊：
影响因子：
--
通讯作者：
Jong
中科院分区：
文献类型：
--
作者：
Artem Lenskiy;Jong
许多车祸都是由于司机的疏忽而发生的。在本文中，我们将提出一种新的系统，用于驾驶员面部和面部特征检测，这可能是驾驶员意识控制的初步步骤。该系统主要包括以下几个步骤：肤色分割、面部特征分割和虹膜位置定位。对于皮肤分割，我们应用多层感知器来近似肤色像素出现的统计概率，并过滤掉概率较小的像素。主要的皮肤片段被分割成以下类别：眼睛，嘴巴，眉毛，和主要的面部区域。为此，我们提出了一种新的分割技术的基础上的非参数估计的概率密度函数（PDF）使用显着的功能。根据从四个PDF中选择的最高概率对每个像素进行分类。检测到的眼睛区域进行分析与圆形霍夫变换，以找到他们的位置和半径。我们测试了我们的系统在真实的enivorment使用CCD相机。该系统的检测率为92%。
Many car accidents happen due to driver inattention. In this paper we will present a novel system for driver facial and facial features detection, which could be a preliminary step toward driver awareness control. This system consists of the following major steps: Skin-color segmentation, facial features segmentation, and iris position location. For skin-segmentation we applied a multi-layer perceptron to approximate the statistical probability of skin-color pixels appearance and filter out pixels with small probabilities. There maining skin-segments were segmented into the following categories: eye, mouth, eyebrow, and there maining facial regions. For this purpose we propose a novel segmentation technique based on non parametric estimation of probability density functions(PDF) using salient features. Each pixelis classified according to the highest probability selected from four PDFs. Detected eye regions are analyzed with the circular Hough transform to find the position and radius of their is. We tested our system in real enivorment using a CCD camera. The system showed a 92% detection rate.