The Gaussian Blue Noise algorithm treats each point as a Gaussian kernel, and optimizes the positions of points so that the density field produced by superimposing the Gaussian kernels approximates the target image density distribution while maintaining blue noise spectral characteristics.