Generation of Heterogeneous PET Images from Uniform Organ Activity Maps Using a Pretrained Domain-Adapted Diffusion Model
A new study presents a pretrained domain-adapted diffusion model for generating heterogeneous PET images from uniform organ activity maps. This model aims to improve the efficiency and accuracy of synthetic PET image generation, which is crucial for various imaging workflows. The results indicate that the generated images closely resemble actual PET images, demonstrating high quantitative accuracy and effective tumor segmentation performance.
- ▪The study developed a pretrained domain-adapted diffusion model for anatomy-conditioned PET synthesis.
- ▪Generated images achieved high quantitative accuracy with concordance correlation coefficients above 0.92.
- ▪The synthesized images showed noise levels and texture characteristics similar to target PET images.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Computer Vision and Pattern Recognition arXiv:2605.20267 (cs) [Submitted on 18 May 2026] Title:Generation of Heterogeneous PET Images from Uniform Organ Activity Maps Using a Pretrained Domain-Adapted Diffusion Model Authors:Suya Li, Kaushik Dutta, Debojyoti Pal, Jingqin Luo, Kooresh I. Shoghi View a PDF of the paper titled Generation of Heterogeneous PET Images from Uniform Organ Activity Maps Using a Pretrained Domain-Adapted Diffusion Model, by Suya Li and 4 other authors View PDF Abstract:Synthetic PET images are valuable for quantitative imaging workflow development, scalable virtual imaging trials, and deep learning model training, but conventional physics-based simulation approaches are computationally intensive, limited in anatomical variability, and often fail…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv cs.AI.
Discussion
0 commentsMore from arXiv cs.AI
