Stair Blue Noise Sampling

摘要

A common solution to reducing visible aliasing artifacts in imagereconstruction is to employ sampling patterns with a blue noisepower spectrum. These sampling patterns can prevent discernibleartifacts by replacing them with incoherent noise. Here, we proposea new family of blue noise distributions, Stair blue noise, which ismathematically tractable and enables parameter optimization to obtainthe optimal sampling distribution. Furthermore, for a givensample budget, the proposed blue noise distribution achieves a significantlylarger alias-free low-frequency region compared to existingapproaches, without introducing visible artifacts in the midfrequencies.We also develop a new sample synthesis algorithm thatbenefits from the use of an unbiased spatial statistics estimator andefficient optimization strategies.