Share Email Print

Proceedings Paper

Modified distance transform with raster scanning value propagation
Author(s): Oleg G. Okun; Sergey V. Ablameyko
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

A new algorithm of the non-Euclidean distance transform with raster scanning value propagation is developed. In this algorithm, a single access to some analyzed pixels is sufficient to obtain their final values, while double processing of every pixel is necessary in other algorithms, of this type. In addition to a raster image, interval coding is used to speed up a processing. This representation does not take much memory space. An application of the proposed distance transform algorithm for object shape reconstruction from disconnected blobs (text symbols are chosen as an example) is given. As compared to the standard raster scanning algorithm, a speed-up factor of 1.3 - 1.4 is obtained, while reconstruction results are the same in both cases.

Paper Details

Date Published: 30 September 1996
PDF: 10 pages
Proc. SPIE 2826, Vision Geometry V, (30 September 1996); doi: 10.1117/12.251807
Show Author Affiliations
Oleg G. Okun, Institute of Engineering Cybernetics (Belarus)
Sergey V. Ablameyko, Institute of Engineering Cybernetics (Belarus)

Published in SPIE Proceedings Vol. 2826:
Vision Geometry V
Robert A. Melter; Angela Y. Wu; Longin Jan Latecki, Editor(s)

© SPIE. Terms of Use
Back to Top