2015 | OriginalPaper | Chapter
Rotation Invariant Texture Analysis Based on Co-occurrence Matrix and Tsallis Distribution
Authors : Mateus Habermann, Felipe Berla Campos, Elcio Hideiti Shiguemori
Published in: Advances in Swarm and Computational Intelligence
Publisher: Springer International Publishing
Activate our intelligent search to find suitable subject content or patents.
Select sections of text to find matching patents with Artificial Intelligence. powered by
Select sections of text to find additional relevant content using AI-assisted search. powered by
This article addressed some extensions of a texture classifier invariant to rotations. Originally, that classifier is an improvement of the seminal Haralick’s paper in a sense that the former is rotation invariant due to a circular kernel, which encompasses two concentric circles with different radii and then the co-occurrence matrix is formed. It is not considered only pixels falling exactly on the circle, but also others in its vicinity according to a Gaussian scattering. Firstly, 6 attributes are computed from each of the 18 texture patterns, after that texture patterns are rotated and a correct classification, considering Euclidian distance, is sought. The present paper assesses the performance of the afore-mentioned approach with some alterations: Tsallis rather than Gaussian distribution; addition of noise to rotated images before classification; and Principal Components Analysis during the extraction of features.