Utilização de métodos de registros não rígidos para análise de disfunções pélvicas em imagens médicas

• Ano de defesa: 2019
• Autor(a) principal: Costa, Cícero Lima
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Uberlândia; Brasil; Programa de Pós-graduação em Ciência da Computação
• Palavras-chave em Português: Registro de imagens; K-means; Disfunções pélvicas; Defecografia; Computação; Image Registration; Pelvic Dysfunctions; Defecography; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
• Link de acesso: https://repositorio.ufu.br/handle/123456789/26725; http://dx.doi.org/10.14393/ufu.di.2019.2300

Pelvic Dysfunctions represent 10% to 20% of visits to gastrointestinal clinics, affecting mainly women over 50 years old. Dysfunctions can be diagnosed by defecography, a dynamic Magnetic Resonance Imaging (MRI) scan. To identify a dysfunction, the radiologist deduces various static measurements at different times and in distinct maneuvers during the examination. However, interobserver variability may occur, for the same exam, analyzed by different specialists, different reports can be generated. In order to reduce interobserver variability and assist the radiologist in interpreting defecography for the diagnosis of pelvic dysfunction, we propose: automatically propagate bladder and anal canal markings along defecography frames for each maneuver via image registration based on a variational model; calculate the regions displacement during defecation and; classify exams based on displacements and generate a pre-diagnosis. According to an expert, there are tests that do not have dysfunction or have low, medium or high level of severity disfunction. Analysis of the results indicates that the system can assist the physician in the diagnostic process. The proposal was successful in pre-diagnosing bladder dysfunction and had difficulty with anal canal dysfunction.