Evaluation of 3D CG Image Colorization Quality Using Visible Digital Watermarking after Noise Removal Based on Sparse Dictionary Learning

Thus far, we discussed to represent image data whether it is possible or not to represent meaning image how requirement is satisfied from a view of image quality by preparing dictionary in advance, and then using sparse coding that represent a part or all of image data by combination as few as possible from these components. As past our study, we try to experiment for noise removal objectively using six types of noise, sparse dictionary learning. However, in this case, it was not enough to consider color component. Therefore, there is a problem at least from the view of practicality. On the other hand, it is difficult for 3D CG images included multidimensional information to restore color. Therefore, in this paper, first we add noise for 3D CG images, and then we remove noise from noise addition images based on sparse coding. After that, we try to experiment whether it is possible or not to restore processed images after noise removal as color image by carrying out the colorization using visible digital watermarking. Finally, we carried out the image quality assessment including color component before and after image processing, and then we discussed these results.