Modified iterative guided texture filtering algorithm

Abstract

Structure-texture decomposition smoothing has been extensively studied due to its wide range of applications in computational photography and image processing. In this paper, we propose a new structure-texture decomposition algorithm which is based on two fundamental ideas: (1) guidance image and (2) iterative smoothing. The guidance image is generated by mitigating high-frequency oscillatory components in the original image. The result is then incorporated in a new generic iterative framework which makes use of well-known guided edge-reserving filters such as bilateral filter (BF), guided filter (GF), domain transform filter (DTF), and the extended Bayesian model averaging filter (BMA) called guided Bayesian model averaging filter (GBMA) to achieve texture smoothing. We have presented a detailed study of the proposed algorithm including: guidance image generation, an evaluation of the guided edge-preserving filters which are incorporated in the proposed iterative framework, the number of iterations for the proposed iterative structure, and the selection of guided edge-preserving filter. We demonstrate that the proposed method is a flexible and effective tool for a wide range of image editing applications including: image abstraction, color pencil drawing, content-aware image resizing, and texture editing. In particular, the proposed approach has the best performance in structure-texture decomposition for an image with low-contrast features.