An efficient denoising method for color images

The existing denoising algorithms are usually either low-quality and computationally light (e.g., mean filtering) or high-quality and computationally expensive (e.g., nonlocal means). In this paper, we propose an algorithm for color image denoising designed between these two extremes, benefiting from lower sensitivity of the human visual system to color spatial variations. We show that our method achieves a perceptual denoising performance same as a state-of-the-art wavelet-based color denoising method at less than 20% of its computational cost.