We propose an iterative method for color printer characterization. We employ a parametric spectral model for color printers based on cellular YNSN (Yule-Nielsen modified spectral Neugebauer) equations. We find the Yule-Nielsen parameter from least squares regression over a training
set of spectral measurements using the Euclidean distance in spectral distributions space as the error criterion. We iteratively divide the cells in the colorant space until the average prediction error Δ