Title :
Light source estimation from a single image
Author :
Nguyen, R.M.H. ; Minh Ngoc Le
Author_Institution :
Comput. Sci. & Eng., HCMC Univ. of Technol., Ho Chi Minh City, Vietnam
Abstract :
One of the most difficult problems in Augmented Reality is how to ensure that the luminance of virtual objects embedded in a real scene is consistent with the real luminance. This paper shows an approach to identifying the illumination information from a single image of a real scene. Based on the assumption that the object used as a light probe has a globally convex shape, we can employ it to determine the illumination information in terms of the direction and the intensity.
Keywords :
augmented reality; brightness; computer vision; realistic images; shape recognition; augmented reality; computer vision; globally convex shape; illumination information identification; light probe; light source estimation; real luminance; real scene; single image; virtual object; Azimuth; Cameras; Light sources; Lighting; Probes; Shape; Vectors; Computer vision; Lambertian surface; illuminant detection;
Conference_Titel :
Control Automation Robotics & Vision (ICARCV), 2012 12th International Conference on
Conference_Location :
Guangzhou
Print_ISBN :
978-1-4673-1871-6
Electronic_ISBN :
978-1-4673-1870-9
DOI :
10.1109/ICARCV.2012.6485343
