Search

Your search keyword '"Yuto Hirasawa"' showing total 11 results
Start Over Author "Yuto Hirasawa"
11 results on '"Yuto Hirasawa"'

5. Improved viewpoint entropy to evaluate material appearance under various lighting positions

Author

Yuto Hirasawa, Hiroshi Kintou, Shoji Yamamoto, Ryota Domon, and Norimichi Tsumura

Subjects
business.industry, Computer science, Weight coefficient, 02 engineering and technology, Viewing angle, 01 natural sciences, Atomic and Molecular Physics, and Optics, 010309 optics, Computer graphics, 020210 optoelectronics & photonics, 0103 physical sciences, 0202 electrical engineering, electronic engineering, information engineering, Entropy (information theory), Eye tracking, Computer vision, Artificial intelligence, business
Abstract

In this paper, we proposed a useful calculation method to perform the perception-based selection of lighting position that emphasizes the material appearance of CG object. The proposed method is based on the conventional viewpoint entropy which is used to find an appropriate viewing angle. To find an appropriate shot for material appearance, we first identified the important surface that has information of gloss reflection related to material appearance with eye tracking equipment. Next, we modified the equation of viewpoint entropy by adding a weight coefficient for emphasizing at important surface. Since this viewpoint entropy retains the independency for object shape, light direction, and viewing direction, our proposed method can extract only important light position which is most representative of the material appearance. From the results of verification with changing shape and material of CG object, we confirmed that our selection of lighting position accomplishes the fine agreement with subjective evaluation which is imposed on the selection of appropriate scene with emphasis of material appearance.

Published
2019
Full Text
View/download PDF

6. Device-independent graininess reproduction

Author

Shoji Yamamoto, Junki Yoshii, Yuto Hirasawa, and Norimichi Tsumura

Subjects
business.industry, Subjective rating, Process (computing), Regression analysis, 02 engineering and technology, 01 natural sciences, Luminance, Atomic and Molecular Physics, and Optics, 010309 optics, 020210 optoelectronics & photonics, 0103 physical sciences, 0202 electrical engineering, electronic engineering, information engineering, Computer vision, Artificial intelligence, business, Mathematics
Abstract

We evaluated a method for building a psychophysically based model of graininess perception for a device-independent graininess-reproduction system. The model was developed through experiments that explored the relationship between physical parameters of graininess objects, subjective rating, and the maximum luminance of the displays used to present the objects. The graininess model was generated via multiple regression analysis of the parameters and was used to calculate curved surfaces for which graininess perception was equalized. Even if the values of maximum luminance on the display are changed in the model, the value of graininess under the changed luminance is held by changing the physical parameters of graininess generation in the model. We found that the proposed model and process for device-independent graininess reproduction were effective for our adopted displays with various maximum luminances.

Published
2018
Full Text
View/download PDF

8. Viewpoint Entropy for Material Appearance

Author

Norimichi Tsumura, Ryota Domon, Hiroshi Kintou, Yuto Hirasawa, and Shoji Yamamoto

Subjects
010309 optics, 020210 optoelectronics & photonics, 0103 physical sciences, 0202 electrical engineering, electronic engineering, information engineering, Entropy (information theory), 02 engineering and technology, Statistical physics, 01 natural sciences, Mathematics
Published
2016
Full Text
View/download PDF

10. Unglossy to Glossy Image Conversion Using Deep Photo Style Transfer

Author

Norimichi Tsumura, Junki Yoshii, Yuto Hirasawa, Kensuke Fukumoto, Takashi Yamazoe, and Shoji Yamamoto

Subjects
Computer science, Computer graphics (images), Transfer (computing), General Chemistry, Atomic and Molecular Physics, and Optics, Computer Science Applications, Electronic, Optical and Magnetic Materials, Image conversion, Style (sociolinguistics)
Published
2020
Full Text
View/download PDF

11. Unglossy to Glossy Image Conversion Using Deep Photo Style Transfer.

Author

Kensuke Fukumoto, Junki Yoshii, Yuto Hirasawa, Takashi Yamazoe, Shoji Yamamoto, and Norimichi Tsumura

Subjects
LUMINANCE (Photometry), SEMANTIC differential scale, IMAGE
Abstract

In this paper, we propose a method of image conversion that changes unglossy images into glossy ones using deep photo style transfer. This style of transfer is expected to reproduce the desired image with a metallic appearance based on the technique used for texture transfer. A practical challenge was undertaken to create a gold, metallic image by transferring the style images of a gold ingot. To avoid excessive loss of color balance, we applied the YCrCb separation technique and used only the Y component to reproduce a glossy appearance. The luminance and saturation of the style image were altered to investigate the influence on the convergence of appearance because the converted images should be dependent on the contents of the images. The results of transfer were assessed by subjective evaluation using the semantic differential method. A style image with an appropriate amount of change in contrast was found suitable for an appropriate glossy appearance. Moreover, contrast in the style image can be appropriately chosen depending on the contents of the original images. [ABSTRACT FROM AUTHOR]

Published
2020
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results