期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2017
卷号：95
期号：17
页码：4313
出版社：Journal of Theoretical and Applied
摘要：In this paper, we propose a method for color region detection in color images and videos. It is based on Gaussian mixture model (GMM) which is calculated by the expectation-maximization (EM) algorithm. We assume that we know the number of Gaussian components in the reference image, but we do not know it in input images. The proposed our approach is composed of two steps. We first estimate GMM parameters using EM algorithm over a reference image including colors regions of interest (ROI). To construct 2-dimensional GMM in the reference, we consider two chrominance features, CbCr-channel from YCbCr color model. The second step is to detect and segment the color regions by using GMM parameters in input images. We decide the color regions by the posterior probability which is Gaussian distributions calculated by GMM in the reference image. Our method can only detect and segment the colors ROI including the Gaussian components from the input image. The experimental results show that it is very effective to detect the predefined multi-colored regions in images and videos.
关键词：Gaussian mixture model (GMM); expectation maximization (EM); YCbCr color model; Color region detection; Segmentation
