基于耳廓特征的生物识别新技术

耳廓检测和特征点的定位是基于耳廓的身份识别与验证新技术研究中非常重要的一部分.耳廓具有数据采集量少,便于建库,色泽分布一致,不受丰富表情影响等诸多优势.采用快速耳廓检测和定位技术,并基于耳廓图像可逆线性变换的方法,将图像分别通过力场和能量场进行描述,利用测试点在力场中运动最终收敛至图像能量局部最小值处这一个特征,对耳廓图像特征点进行定位.在实验分析过程中,首先解决了耳廓自动检测,并在此基础上定位和提取耳廓特征点,通过力场模型的调整,使得场线较好的收敛于某最近似像素点,解决了力场连续性和数字图像离散性的矛盾,取得满意结果.

BIOMETRICS BASED ON EAR FEATURE

Ear detection and ear feature extraction is an essential part of ear biometric identification technology.Ear has some advantages such as low dimension,convenience of database building,color distribution uniformity and stability.Based on the fast ear detection tech-nology and the method of reversible linear transformation,This paper describes the image by force field and energy field.We locate the ear image feature points according to the character that the field lines will converge toward the local minium value of image energy.In the process of experiments,we carry out the automatic ear detection and ear feature points location and extraction.Moreover,we adjust the force field model to resolve the problem of the conflict between discreteness of image samples and continuum of force field so that ear feature points can be extracted at the closest pixel points and come to a better result.