2016/12/19 by João B. Florindo, Florindo, João B., Odemir Martinez Bruno +1
Computer Science · Mathematics · #Image Retrieval and Classification Techniques #Mathematical Dynamics and Fractals #Image Processing and 3D Reconstruction
paper · pdf · doi:10.48550/arxiv.1612.06435
This work presents a novel descriptor for texture images based on fractal\ngeometry and its application to image analysis. The descriptors are provided by\nestimating the triangular prism fractal dimension under different scales with a\nweight exponential parameter, followed by dimensionality reduction using\nKarhunen-Lo `eve transform. The efficiency of the proposed descriptors is\ntested on two well-known texture data sets, that is, Brodatz and Vistex, both\nfor classification and image retrieval. The novel method is also tested\nconcerning invariances in situations when the textures are rotated or affected\nby Gaussian noise. The obtained results outperform other classical and\nstate-of-the-art descriptors in the literature and demonstrate the power of the\ntriangular descriptors in these tasks, suggesting their use in practical\napplications of image analysis based on texture features.\n