Image synthesis with class‐aware semantic diffusion models for surgical scene segmentation
Healthcare Technology Letters · 2025-01 · 12 (1)
DOI: 10.1049/htl2.70003 ↗Kurz gefasst
Fachartikel in der Zeitschrift Healthcare Technology Letters, online erschienen am 31.01.2025. Verfasst von Yihang Zhou und 3 weiteren Autor:innen. Beteiligte Einrichtung: Hamlyn Centre for Robotic Surgery, Department of Surgery and Cancer Imperial College London London UK. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 26 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Royal Society.
Abstract (Originalsprache)
Surgical scene segmentation is essential for enhancing surgical precision, yet it is frequently compromised by the scarcity and imbalance of available data. To address these challenges, semantic image synthesis methods based on generative adversarial networks and diffusion models have been developed. However, these models often yield non‐diverse images and fail to capture small, critical tissue classes, limiting their effectiveness. In response, a class‐aware semantic diffusion model (CASDM), a novel approach which utilizes segmentation maps as conditions for image synthesis to tackle data scarcity and imbalance is proposed. Novel class‐aware mean squared error and class‐aware self‐perceptual loss functions have been defined to prioritize critical, less visible classes, thereby enhancing image quality and relevance. Furthermore, to the authors' knowledge, they are the first to generate multi‐class segmentation maps using text prompts in a novel fashion to specify their contents. These maps are then used by CASDM to generate surgical scene images, enhancing datasets for training and validating segmentation models. This evaluation assesses both image quality and downstream segmentation performance, demonstrates the strong effectiveness and generalisability of CASDM in producing realistic image‐map pairs, significantly advancing surgical scene segmentation across diverse and challenging datasets.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Hamlyn Centre for Robotic Surgery, Department of Surgery and Cancer Imperial College London London UK.
- Imperial College Healthcare NHS Trust London UK.
ORCID-Profile: Yihang Zhou
Förderung
- Royal Society
Zitieren
APA
Zhou, Y., Towning, R., Awad, Z., & Giannarou, S. (2025). Image synthesis with class‐aware semantic diffusion models for surgical scene segmentation. Healthcare Technology Letters, 12(1). https://doi.org/10.1049/htl2.70003
Vancouver
Zhou Y, Towning R, Awad Z, Giannarou S. Image synthesis with class‐aware semantic diffusion models for surgical scene segmentation. Healthcare Technology Letters. 2025;12(1). doi:10.1049/htl2.70003
BibTeX
@article{Zhou2025,
author = {Zhou, Yihang and Towning, Rebecca and Awad, Zaid and Giannarou, Stamatia},
title = {Image synthesis with class‐aware semantic diffusion models for surgical scene segmentation},
journal = {Healthcare Technology Letters},
year = {2025},
volume = {12},
number = {1},
doi = {10.1049/htl2.70003}
}
EUVIMED