TY - GEN
T1 - Semantic Image Synthesis for Abdominal CT
AU - Zhuang, Yan
AU - Hou, Benjamin
AU - Mathai, Tejas Sudharshan
AU - Mukherjee, Pritam
AU - Kim, Boah
AU - Summers, Ronald M.
N1 - Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.
PY - 2024
Y1 - 2024
N2 - As a new emerging and promising type of generative models, diffusion models have proven to outperform Generative Adversarial Networks (GANs) in multiple tasks, including image synthesis. In this work, we explore semantic image synthesis for abdominal CT using conditional diffusion models, which can be used for downstream applications such as data augmentation. We systematically evaluated the performance of three diffusion models, as well as to other state-of-the-art GAN-based approaches, and studied the different conditioning scenarios for the semantic mask. Experimental results demonstrated that diffusion models were able to synthesize abdominal CT images with better quality. Additionally, encoding the mask and the input separately is more effective than naïve concatenating.
AB - As a new emerging and promising type of generative models, diffusion models have proven to outperform Generative Adversarial Networks (GANs) in multiple tasks, including image synthesis. In this work, we explore semantic image synthesis for abdominal CT using conditional diffusion models, which can be used for downstream applications such as data augmentation. We systematically evaluated the performance of three diffusion models, as well as to other state-of-the-art GAN-based approaches, and studied the different conditioning scenarios for the semantic mask. Experimental results demonstrated that diffusion models were able to synthesize abdominal CT images with better quality. Additionally, encoding the mask and the input separately is more effective than naïve concatenating.
KW - Abdomen
KW - CT
KW - Diffusion model
KW - Semantic Image Synthesis
UR - https://www.scopus.com/pages/publications/85187641530
U2 - 10.1007/978-3-031-53767-7_21
DO - 10.1007/978-3-031-53767-7_21
M3 - Conference contribution
AN - SCOPUS:85187641530
SN - 9783031537660
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 214
EP - 224
BT - Deep Generative Models - Third MICCAI Workshop, DGM4MICCAI 2023, Held in Conjunction with MICCAI 2023, Proceedings
A2 - Mukhopadhyay, Anirban
A2 - Oksuz, Ilkay
A2 - Engelhardt, Sandy
A2 - Zhu, Dajiang
A2 - Yuan, Yixuan
PB - Springer Science and Business Media Deutschland GmbH
T2 - 3rd Workshop on Deep Generative Models for Medical Image Computing and Computer Assisted Intervention, DGM4MICCAI 2023 Held in Conjunction with 26th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2023
Y2 - 8 October 2023 through 12 October 2023
ER -