Título: Uso de Mapas Auto Organizáveis de Kohonen para quantização de imagens de Tomografia por Impedância Elétrica reconstruídas pelo software EIDORS
Autores: Eulálio, Alancássio; Ribeiro, Leonardo R.; Valença, Mêuser J. S.; Santos, Wellington P. dos
Resumo: This article presents the application of Kohonen Self-Organizing Maps (SOM) for the quantization of reconstructed electrical impedance tomographic images generated from EIDORS ( Electrical Impedance and Diffuse Optical tomography Reconstruction Software), an extensible software base for EIT (Electrical Impedance Tomography) researchers. EIDORS uses Finite Elements Method to reconstruct images from electrical potential information acquired from skin electrodes. This paper consists of a proof concept: in order to get better defined EIT image boundaries, we employed SOM networks to generate quantized images versions without the smoothing effect resulting from the numeric solution of Laplace’s Equation. The results were qualitatively evaluated, seeming interesting for potential use at the decision support in clinical visual qualitative analysis.
Palavras-chave: Electrical Impedance Tomography; Kohonen Self-Organizing Maps; Image Reconstruction Software; Anti-Smoothing Techniques
Páginas: 7
Código DOI: 10.21528/CBIC2011-30.6
Artigo em pdf: st_30.6.pdf
Arquivo BibTex: st_30.6.bib