Robust Inhibition-Augmented Operator for Delineation of Curvilinear Structures

Abstract

Delineation of curvilinear structures in images is an important basic step of several image processing applications, such as segmentation of roads or rivers in aerial images, vessels or staining membranes in medical images, and cracks in pavements and roads, among others. Existing methods suffer from insufficient robustness to noise. In this paper, we propose a novel operator for the detection of curvilinear structures in images, which we demonstrate to be robust to various types of noise and effective in several applications. We call it RUSTICO, which stands for RobUST Inhibition-augmented Curvilinear Operator. It is inspired by the push-pull inhibition in visual cortex and takes as input the responses of two trainable B-COSFIRE filters of opposite polarity. The output of RUSTICO consists of a magnitude map and an orientation map. We carried out experiments on a data set of synthetic stimuli with noise drawn from different distributions, as well as on several benchmark data sets of retinal fundus images, crack pavements, and aerial images and a new data set of rose bushes used for automatic gardening. We evaluated the performance of RUSTICO by a metric that considers the structural properties of line networks (connectivity, area, and length) and demonstrated that RUSTICO outperforms many existing methods with high statistical significance. RUSTICO exhibits high robustness to noise and texture.

Segmentation and 3D reconstruction of rose plants from stereoscopic images

“…As seen, all the skeletonization methods achieved scores above 0.3, 0.7 and 0.9 for d = [0, 1, 2] respectively, with the exception of RUSTICO and RUSTICO+seg, which got the lowest results. However, these results are similar to those obtained in the original paper for the branch segmentation task (F 1 = 42.65 for the TB-Roses v2 dataset with a tolerance of 2 pixels [55]). RUSTICO's low performance is caused by their nature of finding curvilinear structures rather than finding a skeleton of a blob.…”

“…• RUSTICO [55]: It presents a brain-inspired inhibition mechanism built using trainable filters called B-COSFIRE [3] to delineate curvilinear structures. The inhibition mechanism introduces robustness to noise given that a gray scale image passes through two processes, an excitatory part and an inhibition component.…”

“…The excitatory component is selective for bright lines on dark backgrounds, while the inhibitory counterpart is selective for wider dark lines on bright backgrounds. Although the paper focuses on a more general task, Strisciuglio et al [55] uses it to segment branches as one of their experiments.…”

“…For this, a 2D skeleton is extracted from the binary image. To find the skeletonization method that suits better to our task, we evaluated five methods: Zhang & Suen [60], Parallel thinning [23], 3D skeletonization [37], Medial axis [9], and RUSTICO [55] (see Section 5.3 for algorithms details and evaluation results). These methods succeed in finding the skeleton of the plant.…”

“…The following five methods were evaluated using this criteria to find the skeletonization method that suits better to our problem: Zhang & Suen [60], Parallel thinning [23], 3D skeletonization [37], Medial axis [9], and RUSTICO [55] (see Appendix B for a description of these methods).…”