Using covariance matrices for unsupervised texture segmentation

In this paper we propose an efficient unsupervised texture segmentation method. We introduce the extension of a state-of-the-art segmentation algorithm, which is exclusively based on color cues, by incorporating texture information. We further show how to use covariance matrices of low level features for texture description which can be efficiently calculated based on integral images. Furthermore, a multi-scale extension allows to provide accurate texture segmentation results. An experimental evaluation on a synthetic texture database and images of the Berkeley image database demonstrate the improved performance of the algorithm.

• Type: Conference
• Year: 2008
• Where: ICPR
• Authors: Horst Bischof, Michael Donoser
• Keywords: Accurate Texture Segmentation, Computer Vision, ICPR 2008, Synthetic Texture Database, Unsupervised Texture Segmentation