A computational methodology for the staging of lung tumors considering geometric descriptors

Lung diseases diagnosis, specifically the presence of lung tumors, is usually performed with the support of radiological techniques. Computed tomography is the most widely used imaging technique to confirm the presence of this disease. When several researchers require identifying the morphology of these tumors, they deal problems related to the poor delimitation of the borders associated with the anatomical structures that compound the lung, Poisson noise, the streak artifact and the non-homogeneity of gray levels that define each object in the chest images. In this paper, a methodology has been presented to identify in which stage (staging) the mentioned tumors are. For this, first, anisotropic diffusion filter and magnitude of the gradient filter are used in order to address the aforementioned problems. Second, a smart operator and the level set lgorithm are used to segment lung tumors. Finally, considering these segmentations, a set of geometric descriptors is obtained, and it allows staging of such tumors to be precisely established, generating results that are in high correspondence with the reference data, linked to the analyzed tagged images.

• Autores: Vera, Miguel; Huérfano, Yoleidy; Bravo, Antonio
• Fecha de publicación: 2020
• Institución: Universidad Simón Bolívar
• Repositorio: Repositorio Digital USB
• Idioma: eng
• OAI Identifier: oai:bonga.unisimon.edu.co:20.500.12442/6837
• Acceso en línea: https://hdl.handle.net/20.500.12442/6837; http://www.revhipertension.com/rlh_3_2020/16_a_omputational_methodology.pdf
• Palabra clave: Computerized tomography; Lung tumors; Segmentation; Geometric descriptors; Tomografía computarizada; Tumores pulmonares; Segmentación; Descriptores geométricos
• Rights: openAccess
• License: Attribution-NonCommercial-NoDerivatives 4.0 Internacional