Your browser doesn't support javascript.
loading
Multiple mask and boundary scoring R-CNN with cGAN data augmentation for bladder tumor segmentation in WLC videos.
Freitas, Nuno R; Vieira, Pedro M; Tinoco, Catarina; Anacleto, Sara; Oliveira, Jorge F; Vaz, A Ismael F; Laguna, M Pilar; Lima, Estêvão; Lima, Carlos S.
Afiliação
  • Freitas NR; CMEMS-UMinho, University of Minho, 4800-058 Guimaraes, Portugal; LABBELS - Associate Laboratory, Guimaraes, Portugal. Electronic address: id8003@alunos.uminho.pt.
  • Vieira PM; CMEMS-UMinho, University of Minho, 4800-058 Guimaraes, Portugal; LABBELS - Associate Laboratory, Guimaraes, Portugal.
  • Tinoco C; Department of Urology, Hospital of Braga, 4710-243 Braga, Portugal.
  • Anacleto S; Department of Urology, Hospital of Braga, 4710-243 Braga, Portugal.
  • Oliveira JF; Instituto de Telecomunicações, University of Aveiro, Campus Universitário de Santiago, 3810-193 Aveiro, Portugal; School of Technology and Management, Polytechnic Institute of Leiria, Morro do Lena, Alto Vieiro, Apartado 4163, 2411-901 Leiria, Portugal. Electronic address: jorge.oliveira@co.it.pt.
  • Vaz AIF; Algoritmi Center, University of Minho, Guimaraes, Portugal. Electronic address: aivaz@dps.uminho.pt.
  • Laguna MP; Department of Urology, Istanbul Medipol University, 34214 Istanbul, Turkey; Department of Biomedical Engineering and Physics, AMC-University of Amsterdam, L0-108, 1105 AZ Amsterdam, the Netherlands. Electronic address: plaguna@medipol.edu.tr.
  • Lima E; Life and Health Sciences Research Institute, University of Minho, Campus Gualtar, 4710-057 Braga, Portugal; Department of Urology, CUF Hospitals, 4100-180 Oporto, Portugal. Electronic address: estevaolima@med.uminho.pt.
  • Lima CS; CMEMS-UMinho, University of Minho, 4800-058 Guimaraes, Portugal; LABBELS - Associate Laboratory, Guimaraes, Portugal. Electronic address: clima@dei.uminho.pt.
Artif Intell Med ; 147: 102723, 2024 01.
Article em En | MEDLINE | ID: mdl-38184356
ABSTRACT
Automatic diagnosis systems capable of handling multiple pathologies are essential in clinical practice. This study focuses on enhancing precise lesion localization, classification and delineation in transurethral resection of bladder tumor (TURBT) to reduce cancer recurrence. Despite deep learning models success, medical applications face challenges like small and limited datasets and poor image characterization, including the absence lack of color/texture modeling. To address these issues, three solutions are proposed (1) an improved texture-constrained version of the pix2pixHD cGAN for data augmentation, addressing the tradeoff of generating high-quality images with enough stochasticity using the Fréchet Inception Distance (FID) measure. (2) Introducing the Multiple Mask and Boundary Scoring R-CNN (MM&BS R-CNN), a new mask sub-net scheme where multiple masks are generated from the different levels of the mask sub-net pipeline, improving segmentation accuracy by including a new scoring module to refine object boundaries. (3) A novel accelerated training strategy based on the SGD optimizer with the second momentum. Experimental results show significant mAP improvements the data generation scheme improves by more than 12 %; MM&BS R-CNN proposed architecture is responsible for an improvement of about 1.25 %, and the training algorithm based on the second-order momentum increases mAP by 2-3 %. The simultaneous use of all three proposals improved the state-of-the-art mAP by 17.44 %.
Assuntos
Palavras-chave

Texto completo: 1 Base de dados: MEDLINE Assunto principal: Neoplasias da Bexiga Urinária / Algoritmos Idioma: En Ano de publicação: 2024 Tipo de documento: Article

Texto completo: 1 Base de dados: MEDLINE Assunto principal: Neoplasias da Bexiga Urinária / Algoritmos Idioma: En Ano de publicação: 2024 Tipo de documento: Article