一种基于人脸皮肤图像的色斑检测算法

A Facial Spots Detection Algorithm Based on Human Face Images

在皮肤病诊断、化妆品功效判定和皮肤评价中都需要精确地检测色斑,本文提出了一种数字图像处理的方法提取皮肤图像中人脸色斑。首先截取包含色斑的皮肤区域,用基于小波变换同态滤波校正法消除光照不均匀的影响,接着用形态学高帽低帽变换和对比度拉伸增强图像的对比度,最后用自适应灰度阈值分割检测色斑,再用形态学开运算和闭运算消除检测后的一些杂散点,得到最终的检测结果。经实验验证,该方法的准确率较高。

Facial spots affect the beauty and health of humankind, detection of skin lesion for diagnosis of dermatosis, decision of ef- ficacy of cosmetic and evaluation of skin are necessary.. The algorithm of extract facial spots with digital image processing is based on human face images in this paper. Firstly, select a peace of facial spots skin area, remove illumination effect with the wavelet based homomorphic filtering, secondly, morphologic Top- hat transformation and gray tensile are used to enhance the image contrast, finally, adaptive gray threshold segmentation is used to detect facial spots, after testing the elimination of some spurious points with the morphological open computing and closing operation. Experiment proved that the method has better correctional results.