DCCMAPI/CCET

Uma banca de QUALIFICAÇÃO de DOUTORADO foi cadastrada pelo programa.
DISCENTE: GILVAN VERAS MAGALHÃES JUNIOR
DATA: 10/07/2023
HORA: 16:00
LOCAL: Sala Prof. Pedro
TÍTULO: XRaySwinGen: Automatic Medical Reporting for X-ray Exams with Multimodal Model
PALAVRAS-CHAVES: Multimodal, Transformers, Medical Report, Computer Vision, Natural Language Processing
PÁGINAS: 36
GRANDE ÁREA: Ciências Exatas e da Terra
ÁREA: Ciência da Computação
SUBÁREA: Metodologia e Técnicas da Computação
ESPECIALIDADE: Sistemas de Informação
RESUMO: Properly identifying diseases on chest x-ray imaging is essential to generat- ing high-quality medical reports. However, the effort required to prepare these documents daily takes a lot of time. Deep learning techniques can help re- duce the radiologists’ workload during this process. Thus, this work presents the XRaySwinGen, a multimodal model for automatically generating textual medical reports from x-ray image exams based on Computer Vision techniques associated with Natural Language Processing techniques. The experimental re- sults obtained using two datasets, IU Chest X-ray (frontal images), and NIH Chest X-ray were promising, with the BLEU metric equal to 0.731, 0.377, and 0.396, respectively.
MEMBROS DA BANCA:
Presidente - 470.564.523-53 - PEDRO DE ALCANTARA DOS SANTOS NETO
Externo à Instituição - JOAO PAULO PORDEUS GOMES - UFC
Externo à Instituição - THIAGO ALEXANDRE SALGUEIRO PARDO - USP