Texture Segmentation Using Active Contours Driven by Local Steerable Features
Oriented filters are useful in many computer vision and image processing tasks. It is a kind of oriented filters, rotated very efficiently by taking a suitable linear combination of basis filters. Local steerable feature (LSF), which captures local structure corresponding to any orientation, is distinctive enough to provide useful identity orientation information. This paper investigates novel LSF-guided active contour approach for texture segmentation. The steerable filters are first employed to produce a new energy image, then, the active contour proposed by Chan-Vese (ACWE) is applied to the energy image. These active contours combine the advantages of both the LSF texture representation and the ACWE model, and yield high quality texture segmentation. The performance of the algorithm on some test images is presented and analyzed.
D. Lee and H. B. Wang, "Texture Segmentation Using Active Contours Driven by Local Steerable Features", Applied Mechanics and Materials, Vols. 130-134, pp. 543-546, 2012