Although humans perceive the luster of objects visually and sensuously on a daily basis, there is much debate as to how to express this numerically in order to obtain an index of luster. Gloss can be intentionally expressed by painting something on the 3DCG object. Therefore, we thought that we could somehow quantify the glossiness of 3D CG objects by studying the relationship between glossiness and paint. In this paper, we set parameters related to glossiness and paint on 3D CG object images in the Shitsukan Perception Standard Problem image dataset, performed texture analysis on these patterns, and discussed the results by classifying evaluation values using a Support Vector Machine (SVM) in relation to glossiness, paint, and image quality.