论文标题
肤色提取中EM和FCM算法之间的比较
Comparison between EM and FCM algorithms in skin tone extraction
论文作者
论文摘要
这项研究旨在研究实施EM和FCM算法以进行肤色提取。通过使用EM和FCM算法对肤色的统计建模，评估了三个众所周知的颜色空间的功能，即RGB，HSV和YCBCR提取皮肤提取的能力。结果表明，当使用颜色矢量的所有三个组件时，使用EM算法利用高斯混合模型使用EM算法进行肤色的参数建模。尽管丢弃YCBCR和HSV颜色空间中的亮度组件，EM算法仍可获得最佳结果。详细比较的结果在结论中进行了解释。
This study aims to investigate implementing EM and FCM algorithms for skin color extraction. The capabilities of three well-known color spaces, namely, RGB, HSV, and YCbCr for skin-tone extraction are assessed by using statistical modeling of skin tones using EM and FCM algorithms. The results show that utilizing a Gaussian mixture model for parametric modeling of skin tones using EM algorithm works well in HSV color space when all three components of the color vector are used. In spite of discarding the luminance components in YCbCr and HSV color spaces, EM algorithm provides the best results. The results of the detailed comparisons are explained in the conclusion.
