Paper
14 April 1993 Skeleton metric: a tool for quantitative shape comparison
Jonathan W. Brandt, V. Ralph Algazi
Author Affiliations +
Proceedings Volume 1906, Character Recognition Technologies; (1993) https://doi.org/10.1117/12.143633
Event: IS&T/SPIE's Symposium on Electronic Imaging: Science and Technology, 1993, San Jose, CA, United States
Abstract
In many document processing applications it is necessary to efficiently measure how accurately one shape matches another. Often, it is also necessary that the measurement technique be invariant to rotation and scaling. In this paper, a natural metric for the skeleton is proposed and applied to quantitative shape comparison. This new skeleton metric can be computed particularly efficiently when the skeleton is described with polygonal arcs -- as with the continuous skeleton representation. It is also straightforward and efficient to normalize the orientation and scale of the objects being compared when using this representation. However, the most significant property of the skeleton metric is that it is an upper-bound for the Hausdorff distance between the two shapes. Thus the skeleton metric can be used to ensure a bound on the maximum deviation of the shape boundaries from one another. Using this metric, it becomes possible to introduce simplifying approximations in the skeleton while controlling the error of the corresponding regenerated shape. Thus, the skeleton metric provides a bridge between the qualitative, shape-characterizing aspect of the skeleton and the quantitative, comparative aspect of the Hausdorff metric.
© (1993) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Jonathan W. Brandt and V. Ralph Algazi "Skeleton metric: a tool for quantitative shape comparison", Proc. SPIE 1906, Character Recognition Technologies, (14 April 1993); https://doi.org/10.1117/12.143633
Advertisement
Advertisement
RIGHTS & PERMISSIONS
Get copyright permission  Get copyright permission on Copyright Marketplace
KEYWORDS
Optical character recognition

Error analysis

Distance measurement

Detection and tracking algorithms

Polonium

Bridges

Image processing

Back to Top