Researchers have developed a novel multi-agent system that leverages vision-language models (VLMs) and large language models (LLMs) to automate and enhance the denoising of Positron Emission Tomography (PET) images. This framework aims to replicate expert workflows by dynamically assessing image quality and lesion status, autonomously selecting optimal denoising models and parameters, and incorporating rollback mechanisms for closed-loop feedback. Experiments on Siemens Biograph Vision Quadra PET/CT data demonstrated that the proposed system outperforms traditional methods like U-Net, GAN, and DDPM in improving PET image quality. AI
IMPACT This research could lead to more automated and accurate medical imaging analysis, improving diagnostic capabilities.
RANK_REASON The cluster contains an academic paper detailing a new method for image processing using AI. [lever_c_demoted from research: ic=1 ai=1.0]
- Denoising Diffusion Probabilistic Models
- GAN
- large language models
- Positron emission tomography
- Siemens Biograph Vision Quadra
- U-Net
- vision-language model
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
