Translator Disclaimer
1 August 1991 Salient contour extraction for target recognition
Author Affiliations +
To achieve robust and efficient object recognition, particularly from real outdoor images, methods to reduce clutter and extract salient information of objects need to be developed. Towards this end, the authors present a technique to rank and extract salient contours from a 2-D image acquired by a passive sensor. The goal is to find important contours corresponding to possible objects. The method presented starts with edgels from an edge detector and assigns a saliency measure to linked edgels (contours) based on length, smoothness, and contrast. For length the authors use the number of edgels in the contour; for smoothness they use average change of curvature; and for contrast, the edge magnitude. Contours are ranked by saliency and the more salient contours selected. This method is tested on several real outdoor images of objects in cluttered and occluded conditions. Excellent results are obtained. Performance of this technique is evaluated in the context of a recognition system that matches 2-D image corners with 3-D model vertices. Graphs, using corners on the object of interest and clutter are used to demonstrate the appropriateness of saliency ranking. Curves are plotted to display the percentage of object corners to all image corners for the top few salient contours. Extracting the more salient contours increases the ratio of image corners on the object to all image corners, reducing the search space for the corner matching step in recognition.
© (1991) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Kashi Rao and James H. Liou "Salient contour extraction for target recognition", Proc. SPIE 1482, Acquisition, Tracking, and Pointing V, (1 August 1991);

Back to Top