Multiple mask and boundary scoring R-CNN with cGAN data augmentation for bladder tumor segmentation in WLC videos

Basit Öğe Kaydı
dc.authorid0000-0003-0906-4417
dc.contributor.authorFreitas, Nuno R.
dc.contributor.authorVieira, Pedro M.
dc.contributor.authorTinoco, Catarina
dc.contributor.authorAnacleto, Sara
dc.contributor.authorOliveira, Jorge F.
dc.contributor.authorVaz, A. Ismael F.
dc.contributor.authordel Pilar Laguna Pes, Maria
dc.contributor.authorLima, Estêvão
dc.contributor.authorLima, Carlos S.
dc.date.accessioned2023-12-12T07:08:09Z
dc.date.available2023-12-12T07:08:09Z
dc.date.issued2024
dc.departmentİstanbul Medipol Üniversitesi, Tıp Fakültesi, Cerrahi Tıp Bilimleri Bölümü, Üroloji Ana Bilim Dalı
dc.departmentİstanbul Medipol Üniversitesi, Uluslararası Tıp Fakültesi, Cerrahi Tıp Bilimleri Bölümü, Üroloji Ana Bilim Dalı
dc.description.abstractAutomatic 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 %.
dc.description.sponsorshipFundação para a Ciência e a Tecnologia ; Ministério da Ciência, Tecnologia e Ensino Superioren_US
dc.identifier.citationFreitas, N. R., Vieira, P. M., Tinoco, C., Anacleto, S., Oliveira, J. F., Vaz, A. I. F. ... Lima, C. S. (2024). Multiple mask and boundary scoring R-CNN with cGAN data augmentation for bladder tumor segmentation in WLC videos. Artificial Intelligence in Medicine, 147. https://dx.doi.org/10.1016/j.artmed.2023.102723
dc.identifier.doi10.1016/j.artmed.2023.102723
dc.identifier.issn0933-3657
dc.identifier.scopus2-s2.0-85178106209
dc.identifier.scopusqualityQ1
dc.identifier.urihttps://dx.doi.org/10.1016/j.artmed.2023.102723
dc.identifier.urihttps://hdl.handle.net/20.500.12511/11982
dc.identifier.volume147
dc.identifier.wos001132967000001en_US
dc.identifier.wosqualityQ1
dc.indekslendigikaynakWeb of Science
dc.indekslendigikaynakScopus
dc.indekslendigikaynakPubMed
dc.institutionauthordel Pilar Laguna Pes, Maria
dc.language.isoen
dc.publisherElsevier B.V.
dc.relation.ispartofArtificial Intelligence in Medicineen_US
dc.relation.publicationcategoryMakale - Uluslararası Hakemli Dergi - Kurum Öğretim Elemanı
dc.rightsinfo:eu-repo/semantics/embargoedAccess
dc.subjectLesion Localization
dc.subjectMulti-Pathology Detection
dc.subjectTexture-Constrained GAN
dc.titleMultiple mask and boundary scoring R-CNN with cGAN data augmentation for bladder tumor segmentation in WLC videos
dc.typeArticle

Dosyalar

Orijinal paket
Listeleniyor 1 - 1 / 1
Küçük Resim Yok
İsim:
Laguna-Pilar-2024.pdf
Boyut:
2 MB
Biçim:
Adobe Portable Document Format
Açıklama:
Tam Metin / Full Text
İndir
Lisans paketi
Listeleniyor 1 - 1 / 1
Küçük Resim Yok
İsim:
license.txt
Boyut:
1.44 KB
Biçim:
Item-specific license agreed upon to submission
Açıklama:
İndir

Koleksiyon

Makale Koleksiyonu
Makale Koleksiyonu
PubMed İndeksli Yayınlar Koleksiyonu
Scopus İndeksli Yayınlar Koleksiyonu
WoS İndeksli Yayınlar Koleksiyonu