In real world no color exists. Only spectral light distributions interact to form the final color sensation. This paper presents preliminary experiments whose purpose is to test the robustness of a spatial color computation in relation to changes in the acquisition of spectral information. The basic idea is that human vision system has evolved into a robust system to acquire visual information, in this case the color, adapting to varying illumination conditions to guarantee color constancy. The presented experiments test changes in the output of a Retinex-derived tone mapping operator, varying illuminants and color matching function curves. Synthetic high dynamic range multispectral images have been computed by a photometric ray tracer using different illuminants. Then, using standard and modified color matching functions, a set of high dynamic range RGB images has been created. This set has been converted to standard RGB images using a linear tone mapping algorithm with no spatial color computation and one based on Retinex, performing a spatial color normalization. A discussion of the results is presented.
|Titolo:||Spectral information and spatial color computation|
|Autori interni:||RIZZI, ALESSANDRO (Primo)|
GADIA, DAVIDE (Secondo)
MARINI, DANIELE LUIGI ROBERTO (Ultimo)
|Settore Scientifico Disciplinare:||Settore INF/01 - Informatica|
|Data di pubblicazione:||2005|
|Digital Object Identifier (DOI):||10.1117/12.585896|
|Tipologia:||Book Part (author)|
|Appare nelle tipologie:||03 - Contributo in volume|
File in questo prodotto:
- PubMed Central loading...