#56: Adaptive robust structure tensors for orientation estimation and image segmentation

S. Nath and K. Palaniappan

Lecture Notes in Computer Science (ISVC), Volume 3804, pgs. 445--453, 2005

image enhancement, features, detection, segmentation, denoising, convex relaxation

PlainText, Bibtex, PDF, DOI, Google Scholar


Recently, Van Den Boomgaard and Van De Weijer have pre- sented an algorithm for texture analysis using robust tensor-based esti- mation of orientation. Structure tensors are a useful tool for reliably esti- mating oriented structures within a neighborhood and in the presence of noise. In this paper, we extend their work by using the Geman-McClure robust error function and, developing a novel iterative scheme that adap- tively and simultaneously, changes the size, orientation and weighting of the neighborhood used to estimate the local structure tensor. The itera- tive neighborhood adaptation is initialized using the total least-squares solution for the gradient using a relatively large isotropic neighborhood. Combining our novel region adaptation algorithm, with a robust tensor formulation leads to better localization of low-level edge and junction image structures in the presence of noise. Preliminary results, using syn- thetic and biological images are presented.