Abstract-We propose shiny analysis framework accompanied with makeup deterioration using normalized facial images (MaVIC and the corresponding makeup-deteriorated data sets). These images are analyzed and reconstructed based on principal component analysis (PCA) and then the differential ones between the reconstruction images with different numbers of PCA components can be generated. The shiny Eigen-faces contributing to large shiny transition can automatically be selected according to the total variances of the difference images, and then variances of the reconstructed images with only the shiny Eigen-faces can be used as measurement of shine degree. In addition, we execute PCA analysis in several color spaces, and explore the possibility of the color space, which can best manifest the shine existing in the facial images.