A Composite Model for Representing Spectral Functions

In this report we propose a new model to represent spect ral functions called the composite model. This model is built on the idea that we decompose all spe ctral functions into smooth and spiky components, each with its own distinct representation. Specifically, a smooth spectrum can be expressed as a linear combination of a set of given basis functions and thus be represented through the corresponding coefficients. The discrete spike s in a spiky spectral function can be represented directly through their locations and heights. For the smooth spectra, we propose re- sampling the functions that are reconstructed from the coefficients in the linear combination to improve efficiency. Spectral multiplications are thus greatly reduced in complexity. This new model demonstrates remarkable advantages in representing spectral functions with aspect to accuracy, compactness, computational efficiency, portabil ity, and flexibility. We expect that this work will significantly contribute to the research and applications in color science, realistic image synthesis, and color image analysis.