Extração de atributos e classificação de lesões em imagens colposcópicas na prevenção do câncer do trato genital inferior

• Ano de defesa: 2006
• Autor(a) principal: Mussoi, Susana Rosa
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• dARK ID: ark:/26339/001300000r1vb
• Idioma: por
• Instituição de defesa: Universidade Federal de Santa Maria; BR; Engenharia de Produção; UFSM; Programa de Pós-Graduação em Engenharia de Produção
• Palavras-chave em Português: Processamento de imagens; Segmentação de imagens colposcópicas; Câncer genital; Colposcopia; Images processing; Colposcopy images segmentation; Cervical cancer; Colposcopy; CNPQ::ENGENHARIAS::ENGENHARIA DE PRODUCAO
• Link de acesso: http://repositorio.ufsm.br/handle/1/8024

Cervical cancer is a pathology that can be prevented through the diagnosis of pre-invasive alterations. These initial lesions can, in these periods, be submitted to some types of individualized treatments, avoiding the malignant neoplasia that, in general, it is invasive and lethal. The development of new methods for diagnosis is necessary. This work considers a way of images processing through tools of segmentation of digital colposcopy images. Morphological operators were applied to segment acetowhite and mosaic image data. The proposed methodology aimed at to assist in the identification of cervical pre-cancer lesions in relation the attributes of form, size, coloration, shading and edges that they can be easily accepted for different programs, being able to be used as plus a subsidy for the investigation and diagnosis of pathologies of the inferior genital region, assisting professionals of the health area that works in this speciality. The toolbox was developed using Mmorph for MatLab.