7th IMRMPT

Abstract

Abstract information

English title:	Statistical techniques based on variants of the geometric mean filter applied to 3-D preprocessing of computed tomography medical images.
Spanish title:	T�cnicas estad�sticas basadas en variantes del filtro de media geom�trica aplicadas al preprocesamiento 3-D de im�genes m�dicas de tomograf�a computarizada.
English abstract:	The study, treatment and processing of medical images is of special interest in the clinical and imaging context due to its importance and contributions in the field of medicine and bioengineering with special interest in clinical diagnosis. Likewise, computed tomography (MSCT) is one of the fastest growing imaging modalities in clinical diagnostics. However, the images present imperfections that affect the quality of the information present in the images and hinder the identification and monitoring of relevant pathologies in the issuance of a more accurate diagnosis. To reduce the impact of imperfections, such as noise and artifacts, a preprocessing stage was carried out by applying a set of statistical techniques based on variants of the geometric mean filter, which were compared, from the point of view of efficiency and effectiveness with classical statistical techniques that have been widely reported in medical image preprocessing. During this stage, 02 3-D databases (DB), one synthetic DB and one MSCT DB, were preprocessed. For the validation of the preprocessing techniques, a set of statistical metrics was applied for quantitative and qualitative evaluation by visual inspection of the images. The results showed that the variant of the geometric mean filter that considers the square of the arithmetic mean and the square of the variance obtained the best Score in the preprocessing with respect to the other techniques considered, with a time of 0.83�0.0010 minutes and window size 3x3x3.
Spanish abstract:	El estudio, tratamiento y procesamiento de las im�genes m�dicas, es de especial inter�s en el contexto cl�nico y de imagenolog�a por su importancia y contribuciones en el �mbito de la medicina y la Bioingenier�a con en especial inter�s en el diagn�stico cl�nico. As� mismo la tomograf�a computarizada (MSCT) es una de las modalidades de adquisici�n de im�genes de mayor crecimiento en los diagn�sticos cl�nicos. Sin embargo, las im�genes presentan imperfecciones que afecta la calidad de la informaci�n presente en las im�genes y que dificultan la identificaci�n y monitoreo de patolog�as relevantes en la emisi�n de un diagn�stico m�s preciso. Para disminuir el impacto de las imperfecciones, como ruidos y artefactos, se llev� a cabo una etapa de preprocesamiento mediante la aplicaci�n de un conjunto de t�cnicas estad�sticas basadas en variantes del filtro de media geom�trica, las cuales fueron comparadas, desde el punto de vista de eficiencia y efectividad con t�cnicas estad�sticas cl�sicas que han sido ampliamente reportadas en preprocesamiento de im�genes m�dicas. Durante esta etapa se preprocesaron 02 bases de datos (BD) 3-D, una BD sint�tica y una BD de MSCT. Para la validaci�n de las t�cnicas de preprocesamiento se aplic� un conjunto de m�tricas estad�sticas para la evaluaci�n cuantitativa y cualitativamente por inspecci�n visual de las im�genes. Los resultaros mostraron que la variante del filtro de media geom�trica que considera el cuadrado de la media aritm�tica y el cuadrado de la varianza obtuvo el mejor Score en el preprocesamiento respecto a las otras t�cnicas consideradas, con un tiempo de 0.83�0.0010 minutos y tama�o de ventana 3x3x3.
Topic:
Presentation Type:
Speaker:

Authors

Name	Institution	Email
Oscar Valbuena Prada	Universidad Sim�n Bol�var - sede C�cuta	oscar.valbuena@unisimon.edu.co

Presentation

Room:	Principal Room
Date:	June 8, 2023
Hour:	18:00 - 18:20

Back

7^th IMRMPT
Seventh International Meeting for Researchers in Materials and Plasma Technology

Previous editions:

        1st IMRMPT 2011
        2nd IMRMPT 2013
        3rd IMRMPT 2015
        4th IMRMPT 2017
        5th IMRMPT 2019
        5+1 IMRMPT 2021

Contact information:
imrmpt@foristom.org

Copyright 2022 FORISTOM
All rights reserved