Multiple view image denoising using 3D focus image stacks

In this paper, we propose an improved fast multiple-view image denoising algorithm based on plane sweeping (PS) and non-local means (NLM). Miyata et al. proposed to use PS to convert multi-view images into multi-focus images for noise suppression. It showed improved computational efficiency and comparable denoising quality compared to conventional methods. In this paper, we introduce the notion of 3D focus image stacks which we called super images, and use an adaptation of NLM algorithm to further improve the performance of multi-view denoising. Experiments show that our improved algorithm outperforms that reported by Miyata et al. with slight increase of computation time, and is much faster than conventional NLM denoising method.