Segmenting textured 3D surfaces using the space/frequency representation.

Abstract

Segmenting 3D textured surfaces is critical for general image understanding. Unfortunately, current efforts of automatically understanding image texture are based on assumptions that make this goal impossible. Texture-segmentation research usually assumes that the textures are flat and viewed from the front, while shape-from-texture work assumes that the textures have already been segmented. This deadlock means that none of these algorithms will work reliably on images of 3D textured surfaces. An algorithm has been developed by the authors that can segment an image containing nonfrontally viewed, planar, periodic textures. The spectrogram (local power spectrum) is used to compute local surface normals from small regions of the image. This algorithm does not require unreliable image feature detection. Based on these surface normals, a 'frontalized' version of the local power spectrum is computed that shows what the region's power spectrum would look like if viewed from the front. If neighboring regions have similar frontalized power spectra, they are merged. The merge criterion is based on a description length formula. The segmentation is demonstrated on images with real textures. To the authors' knowledge, this is the first program that can segment 3D textured surfaces by explicitly accounting for 3D shape effects.