Back to articles
Articles
Volume: 26 | Article ID: art00025
Image
Converting the Images without Glossiness into the Images with Glossiness by using Deep Photo Style Transfer
  DOI :  10.2352/ISSN.2169-2629.2018.26.145  Published OnlineNovember 2018
Abstract

In this paper, we propose an image conversion method to transfer the images without glossiness into the images with glossiness by using deep photo style transfer technique. The deep style photo transfer can be expected to reproduce a desired image with metallic appearance based on the texture transfer technique. Our practical challenge was performed to create the gold metallic image by transferring a style image of the gold ingot. Two kinds of stile images where one gold ingot and assembled mass of gold ingots were tested to verify how degree of complexity in style image is appropriate for our propose using deep neural network. In order to avoid an excessive loss of color balance, we also applied the YCrCb separation technique and used only Y component to learn the only style of gloss appearance. Moreover, the luminance and saturation of the style image were changed to investigate the influence into the converted appearance, since the converted appearances are expected to have the dependence with the contents of the images. These transferred results by changing the luminance and saturation of style image were evaluated by subjective evaluation using semantic differential method. As the results, it is found that the style image with an appropriate amount of contrast change is suitable for appropriate gloss appearance, then showed that there is appropriate selection of contrast in style image depending the contents of original images.

Subject Areas :
Views 9
Downloads 1
 articleview.views 9
 articleview.downloads 1
  Cite this article 

Kensuke Fukumoto, Junki Yoshii, Yuto Hirasawa, Takashi Yamazoe, Shoji Yamamoto, Norimichi Tsumura, "Converting the Images without Glossiness into the Images with Glossiness by using Deep Photo Style Transferin Proc. IS&T 26th Color and Imaging Conf.,  2018,  pp 145 - 150,  https://doi.org/10.2352/ISSN.2169-2629.2018.26.145

 Copy citation
  Copyright statement 
Copyright © Society for Imaging Science and Technology 2018
72010350
Color and Imaging Conference
color imaging conf
2166-9635
Society for Imaging Science and Technology