Color Spectrum Estimation from RGB Triplet by Using Photon Statistics
直接从RGB估计光谱的量子统计方法

The RGB triplet of a pixel in a color image contains more information than itself. This paper proposes a novel approach for color spectrum estimation directly from the corresponding RGB signals. (more specific, the resultant spectrum is a metamer of the given RGB). It stems from the idea of optical quantum statistics which regards a spectrum as the random distribution of many of the photons. After deriving a maximum probably equation, we transform the estimating problem to an optimization problem. Experiments on real images show the practicable of the proposed novel method. We believe that it provides many new tools for image processing and computer graphics that involve color information.