Share Email Print
cover

Proceedings Paper

Generation of realistic scene using illuminant estimation and mixed chromatic adaptation
Author(s): Jae-Chul Kim; Sang-Gi Hong; Dong-Ho Kim; Jong-Hyun Park
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

The algorithm of combining a real image with a virtual model was proposed to increase the reality of synthesized images. Currently, synthesizing a real image with a virtual model facilitated the surface reflection model and various geometric techniques. In the current methods, the characteristics of various illuminants in the real image are not sufficiently considered. In addition, despite the chromatic adaptation plays a vital role for accommodating different illuminants in the two media viewing conditions, it is not taken into account in the existing methods. Thus, it is hardly to get high-quality synthesized images. In this paper, we proposed the two-phase image synthesis algorithm. First, the surface reflectance of the maximum high-light region (MHR) was estimated using the three eigenvectors obtained from the principal component analysis (PCA) applied to the surface reflectances of 1269 Munsell samples. The combined spectral value, i.e., the product of surface reflectance and the spectral power distributions (SPDs) of an illuminant, of MHR was then estimated using the three eigenvectors obtained from PCA applied to the products of surface reflectances of Munsell 1269 samples and the SPDs of four CIE Standard Illuminants (A, C, D50, D65). By dividing the average combined spectral values of MHR by the average surface reflectances of MHR, we could estimate the illuminant of a real image. Second, the mixed chromatic adaptation (S-LMS) using an estimated and an external illuminants was applied to the virtual-model image. For evaluating the proposed algorithm, experiments with synthetic and real scenes were performed. It was shown that the proposed method was effective in synthesizing the real and the virtual scenes under various illuminants.

Paper Details

Date Published: 18 December 2003
PDF: 9 pages
Proc. SPIE 5293, Color Imaging IX: Processing, Hardcopy, and Applications, (18 December 2003); doi: 10.1117/12.524639
Show Author Affiliations
Jae-Chul Kim, Electronics and Telecommunications Research Institute (South Korea)
Sang-Gi Hong, Electronics and Telecommunications Research Institute (South Korea)
Dong-Ho Kim, UPSON Co., Ltd. (South Korea)
Jong-Hyun Park, Electronics and Telecommunications Research Institute (South Korea)


Published in SPIE Proceedings Vol. 5293:
Color Imaging IX: Processing, Hardcopy, and Applications
Reiner Eschbach; Gabriel G. Marcu, Editor(s)

© SPIE. Terms of Use
Back to Top