研究一种红外医学图像处理与分析方法,实现红外人脸图像中特征区域的自动定位。方法

针对红外正面脸部图像,采用一种无监督的局部和全局的特征提取方法,首先通过阈值法区分出前景和

背景,并根据面部特征对称性在前景中确定鼻区 然后在面部确定一个包含所有特征的矩形区域,利用

Harris算子在该区域检测出角点,并找出这些点的局部最大值点 最后用K-means方法对这些点进行

聚类

-To develop an mi age analyzing procedure forautomatic localization of facial fea-

tures on infrared mi ages.M ethodsAn unsupervised localand global features extractionmethodwas

adopted for the localization of facial features of frontalview face mi age. First, a thresholdwas used

to segment the mi age into foreground and background, and the nose was localized by considering

the symmetry of the face. Second, Harris operatorwas adopted to detect interest points in a rec-

tangulararea covering allthe facial features, and then localmaxmi um of the interestpointswere de-

tected. And finally, K-means clusteringmethodwas used to cluster the points and obtain the facial

features localization.