| Authors: | Aktaş, Mehmet Ali Žunić, Joviša |
Affiliations: | Mathematical Institute of the Serbian Academy of Sciences and Arts | Title: | Measuring shape ellipticity | Series/Report no.: | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | First page: | 170 | Last page: | 177 | Related Publication(s): | Computer Analysis of Images and Patterns | Conference: | 14th International Conference, CAIP 2011, Seville, Spain, August 29-31, 2011 | Issue Date: | 20-Sep-2011 | Rank: | M33 | ISBN: | 9783642236716 | ISSN: | 0302-9743 | DOI: | 10.1007/978-3-642-23672-3_21 | Abstract: | A new ellipticity measure is proposed in this paper. The acquired shape descriptor shows how much the shape considered differs from a perfect ellipse. It is invariant to scale, translation, rotation and it is robust to noise and distortions. The new ellipticity measure ranges over (0, 1] and gives 1 if and only if the measured shape is an ellipse. The proposed measure is theoretically well founded, implying that the behaviour of the new measure can be well understand and predicted to some extent, what is always an advantage when select the set of descriptors for a certain application. Several experiments are provided to illustrate the behaviour and performance of the new measure. |
Keywords: | early vision | Shape | shape descriptors | shape ellipticity | Publisher: | Springer Link |
Show full item record
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.