Compressed Sensing for Reconstruction of Reflectance Spectra from Tristimulus Values

The present work deals with the reconstruction of the spectrum from the tristimulus values that is spectrally similar to the original function. It is known that in the estimation of reflectances from a small number of responses of colour sensors, some wavelengths occur with very high probability. Therefore the spectra of reflectances are sparse in nature. The recently proposed Compressed Sensing (CS) is a powerful tool for recovering the sparsest solution of under-determined system of equations. Intending to exploit the sparsity of colour spectrum, this paper attempts to recover the colour spectrum from tristimulus values in the framework of CS. The experimental results presented herein vindicate that the CS based procedure provides better reconstruction than the method based on the standard pseudo inverse.