ESTRO 2024 - Abstract Book

S3905

Physics - Image acquisition and processing

ESTRO 2024

in dual energy X-ray imaging based on Deformable Image Registration," 2019 IEEE Nuclear Science Symposium and Medical Imaging Conference (NSS/MIC), Manchester, UK, 2019, pp. 1-2, doi: 10.1109/NSS/MIC42101.2019.9059964 Acknowledgement: Nawal Alqethami acknowledges the scholarship from the Ministry of Education in Riyadh, Saudi Arabia funded through the Saudi Cultural Bureau in Berlin, Germany. Marco Riboldi acknowledges the support of the Deutsche Forschungsgemeinschaft (DFG Research Instrumentation Program n. INST 86/2120-1 FUGG)

2255

Digital Poster

Iterative CT Reconstruction with Generative AI

Sho Ozaki 1 , Shizuo Kaji 2 , Kanabu Nawa 3 , Toshikazu Imae 3 , Keiichi Nakagawa 3

1 Hirosaki University, Graduate School of Science and Technology, Aomori, Japan. 2 Kyushu University, Institute of Mathematics for Industry, Fukuoka, Japan. 3 The University of Tokyo Hospital, Department of Radiology, Tokyo, Japan

Purpose/Objective:

Generative AI has garnered significant attention in recent years. In particular, diffusion models, which serve as the foundational models for generative AI, such as Stable Diffusion and DALL-E3, can produce high-quality images. The performance of these models surpasses that of GAN-based models in natural image processing. In the field of medicine, there is a substantial demand for improved CT image quality, particularly in applications like cone-beam CT (CBCT) and mega-voltage CT (MVCT), which are utilized in image-guided radiation therapy (IGRT). We have employed the diffusion model to enhance CT image quality.

Material/Methods: