A Novel Smoothing Filter with Texture Suppression for Fabric Images

Abstract

Extracting patterns in fabric images usually should preserve the color of yarns and pattern structures, and suppress textures caused by adjacent yarns. A novel filter is proposed for smoothing fabric images while suppressing textures and preserving fabric pattern structures, by means of a nonlinear combination of local color values and lightness gradients underlying CIE-Lab color space. The proposed filter smooths color values and preserves pattern structures based on photometric similarity of neighboring pixels of the fabric images, and suppresses textures based on local gradient difference between the textures of yarns and the patterns of fabric images. It takes a non-iterative, local, and simple scheme similar as bilateral filter. Compared with popular smoothing filters, the experiments demonstrate that the proposed filter produces satisfactory smoothing images with the properties of texture-suppressing and pattern-preserving.