Abstract
The automatic extraction of elongated curvilinear structures (CLSs) is an important task in various image processing applications, including numerous remote sensing, and biometrical and medical problems. To address this task, we develop a stochastic approach that relies on a fixed-grid, localized Radon transform for line segment extraction and a conditional random field model to incorporate local interactions and refine the extracted CLSs. We propose several different energy data terms, the appropriate choice of which allows us to process images with different noise and geometry properties. The contribution of this paper is the design of a flexible and robust elongated CLS extraction framework that is comparatively fast due to the use of a fixed-grid configuration and fast deterministic Radon-based line detector. We present several different applications of the developed approach, namely: 1) CLS extraction in mammographic images; 2) road networks extraction from optical remotely sensed images; and 3) line extraction from palmprint images. The experimental results demonstrate that the method is fairly robust to CLS curvature and can accurately extract blurred and low-contrast elongated CLS.
Highlights
T HE task of automatic detection of straight lines and, more generally, curvilinear structures (CLS) in images is one of the fundamental problems in image processing and pattern recognition
In some image processing problems CLS detection is a necessary preprocessing stage when complex objects are of interest that comprise distinct combinations of linear features, such as spicule pattern analysis for cancer/mass detection in mammography [4]
We propose a two-step stochastic approach for the extraction of elongated CLS from images of natural and man-made scenes affected by noise, blur, acquisition artifacts and the presence of non-curvilinear structures
Summary
T HE task of automatic detection of straight lines and, more generally, curvilinear structures (CLS) in images is one of the fundamental problems in image processing and pattern recognition. The aim of the first step is to extract a wide set of line segments including the CLS of interest To this end we apply a localized version of the Radon transform to each of the image regions over an overlapping square grid and extract its first several maxima as line segment candidates. The designed extraction technique (i) uses a CRF design on a fixed grid to significantly reduce the MCMC complexity and optimization cost and (ii) forms a flexible framework whereby the selection of distinct application-specific unary terms facilitate CLS extraction in images affected by a variety of noise factors, such as blur, occlusion, low-contrast, and background clutter effects.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
