New VLM framework automates cardiac MRI quality assessment for ablation planning

Researchers have developed a novel two-stage vision language model (VLM) framework to automate the assessment of clinical quality and usability for LGE-MR images used in cardiac ablation planning. The first stage employs a fine-tuned VLM to generate radiology-style reports on criteria like noise, motion artifacts, and boundary accuracy. The second stage uses a GPT-based module to map these reports to structured quality scores and a binary decision on clinical usability. In benchmarking, InternVL2 showed high accuracy in criterion-level assessment, while DeepSeek achieved perfect agreement on clinical usability. AI

IMPACT This research could lead to more consistent and scalable quality assessment of medical imaging, improving the safety and outcomes of cardiac ablation procedures.

RANK_REASON The cluster contains a research paper detailing a new methodology and benchmark results for a specific application. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• Cardiac Ablation Planning
• DeepSeek
• GPT
• InternVL2
• LGE-MR Images
• Radiologists
• Vision Language Model

• paper
• product

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →