Abstract
Most existing line segment detection methods suffer from the over-segmentation phenomenon. An improved line segment detection method is presented in this work, which can generate more and longer line segments, yet still accurately reflect the structural details of the image. Line segment grouping, line segment validation and a multiscale framework are adopted to reach this end. Specifically, smart grouping rules are introduced to locate potential homologous line segments (derived from the same boundaries). Novel merging criteria based on Helmholtz principle is then used to evaluate the meaningfulness between separate line segments and their merged ones. The improved multiscale framework facilitates line segments merging in detection and post-detection processes, yielding more high-quality line segments. Finally, the proposed method is compared with four leading methods on the famous public dataset, YorkUrban-LineSegment. Experimental results demonstrate that the method has good continuity and outperforms the leading methods in F-measure.
Highlights
A line segment (LS) is a common and important geometric primitive in a digital image as most objects in man-made scenes are structured and can be outlined by LSs
LS detection algorithms can be used as an assistant tool for generating image proposals that contain salient straight lines and junctions, feeding these image proposals to convolutional neural network (CNN) for further extracting wireframes of images [8]
A high-quality LS detection method is proposed in this paper which tends to generate more and longer LSs but can still accurately reflect structural information of the image
Summary
A line segment (LS) is a common and important geometric primitive in a digital image as most objects in man-made scenes are structured and can be outlined by LSs. LS detection has been studied extensively and remains an active field within image processing research [7]. LS detection algorithms can be used as an assistant tool for generating image proposals that contain salient straight lines and junctions, feeding these image proposals to convolutional neural network (CNN) for further extracting wireframes of images [8]. This process achieves desirable results and indicates that traditional image processing techniques can work in collaboration with advancing deep learning methods.
Sign-in/Register to view highlights & summary 
Published Version (
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