Computer Science · Computer Vision and Pattern Recognition
Monte Carlo Noise Reduction Using Bayesian Method in Wavelet domain
E. Trucco, M. Chantler (editors, Ruifeng Xu, Sumanta N. Pattanaik
See on the mapAbstract
A novel post-processing approach for removing Monte Carlo noises in synthetic images is presented in this paper. This paper first presents our findings on the statistical characteristics of the Monte Carlo noise, and then proposes a Bayesian method to remove this noise. The aim of this approach is to efficiently produce high quality synthetic images using Monte Carlo based rendering at low sampling rates.