Leonardo R. Ribeiro,  Mêuser J. S. Valença,  Wellington P. Dos Santos. 

This article presents a quantitative comparison among four grouping algorithms - K-SOM, K-means, K-means optimized with evolutionary programming and Fuzzy C-means - all applied to reconstructed electrical impedance tomographic images generated from EIDORS (Electrical Impedance and Diffuse Optical Tomography Reconstruction Software). EIDORS is an extensible software base for EIT (Electrical Impedance Tomography) researchers. It solves the main equation of EIT, the Laplace's equation, resulting in reconstructed smooth images with no definition of boundaries. Grouping algorithms are used to quantize the resulting reconstructed images from EIDORS in order to get better defined EIT images boundaries. To compare the grouping algorithms we used the Omran combined index that unites three of the main criteria used to validate the best grouping process.

http://www.lbd.dcc.ufmg.br/colecoes/cbsf/2012/0056.pdf

Processando consulta.