摘要

Statistics of natural images have shown to be non-Gaussian quantities displaying high kurtosis, heavy tails and sharp central cusps, i.e. sparse distributions. In this paper we extend on these results to show that the non-linear filter response distributions of natural colour images are also sparse. We then incorporate the statistics of natural images into Bayesian models to illustrate image processing applications in colorization of gray images, compression of colour images and reconstruction of chroma channels from corrupted data.

  • 出版日期2011-9