An optimized set of 3D fractal and multifractal features for the epileptogenic focus characterization in SPECT imaging

Authors:
R. Lopes;M. Vermandel;A. S. Dewalle-Vignion;S. Maouche;N. Betrouni
Affiliations:
Inserm, Pavillon Vancostenobel, CHRU Lille, Lille Cedex, France and Biophysics Lab, UPRES EA 1049, University of Lille 2, Lille, France and LAGIS, CNRS UMR, USTL, Villeneuve d'Ascq, France;Inserm, Pavillon Vancostenobel, CHRU Lille, Lille Cedex, France and Biophysics Lab, UPRES EA 1049, University of Lille 2, Lille, France and Department of Nuclear Medicine, University Hospital of L ...;Inserm, Pavillon Vancostenobel, CHRU Lille, Lille Cedex, France and Biophysics Lab, UPRES EA 1049, University of Lille 2, Lille, France and Department of Nuclear Medicine, University Hospital of L ...;LAGIS, CNRS UMR, USTL, Villeneuve d'Ascq, France;Inserm, Pavillon Vancostenobel, CHRU Lille, Lille Cedex, France
Venue:
ISBI'09 Proceedings of the Sixth IEEE international conference on Symposium on Biomedical Imaging: From Nano to Macro
Year:
2009

Citing 4
Cited 0

Image interpolation and resampling

Handbook of medical imaging
Near optimum estimation of local fractal dimension for image segmentation

Pattern Recognition Letters
Pattern Classification (2nd Edition)

Pattern Classification (2nd Edition)
Multiclass SVM-RFE for product form feature selection

Expert Systems with Applications: An International Journal

Quantified Score

Hi-index	0.00

Visualization

Abstract

Fractal geometry may be an efficient tool for texture analysis in medical imaging. However its application is primarily restricted to 2D cases and at the only use of an approximation method of the fractal dimension (FD). Recently, multifractal analysis has showed interesting results in this field. This study focuses on the use of an optimized set of 3D fractal and multifractal features for the epileptogenic focus characterization in SPECT imaging. Our results showed that this optimized set, compared to various texture features, improved the classification rate by Support Vector Machines (SVM). Moreover, results were significantly better than the clinical method: SISCOM (Substraction Ictal SPECT Co-registred to MRI).