Arquitetura para recuperação de imagens diagnósticas baseada em conteúdo: uma ferramenta para auxílio à radiologia em ambiente PACS

• Ano de defesa: 2012
• Autor(a) principal: Berni, Cristiano Albiero
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Santa Maria; BR; Ciência da Computação; UFSM; Programa de Pós-Graduação em Informática
• Palavras-chave em Português: CBIR; PACS; DICOM SR; Radiologia; Recuperação de imagens diagnósticas; Busca baseada em conteúdo; Radiology; Diagnostic image retrieval; Content-based search; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
• Link de acesso: http://repositorio.ufsm.br/handle/1/5410

One of the main forms of diagnosis used nowadays matches the exams performed by analysis of diagnostic images. Due to a growing request for this kind of diagnostic and the repetitive manual procedure of the used methods by radiologists new ways are emerging to aid procedures. A tool that can help the physician to report a diagnosis is searching similar cases for that which is being held with the main function of increased safety to the radiologist in his notes. For this, a modular architecture for content-based diagnostic image retrieval was developed as a tool to aid diagnosis. Through the DICOM SR standard used to store radiological findings and measurements - commonly from CAD - was implemented in a PACS environment a structure that will provide storage and query contents extracted from diagnostic images. The contents extraction from images can be done by different processing methods that generate different parameters for storage and retrieval. The project was developed in partnership with a provider of solutions for PACS and the Applied Computing Laboratory of the Federal University of Santa Maria.