Researchers have developed three automated methods to distinguish between intimal and medial intracranial arterial calcifications from CT head scans. These approaches, including an adaptation of radiological visual scores, a sphericity-based method, and a shape foundation model, achieved comparable performance. The shape embedding method showed the best results, with a weighted F1 score of up to 71.5% for single artery classification and 59.8% for joint artery classification, demonstrating robustness even with automated segmentation masks. AI
IMPACT This research demonstrates the feasibility of using AI for precise medical image analysis, potentially aiding radiologists in diagnosis and prognosis.
RANK_REASON The item is an academic paper detailing a new method for medical image analysis. [lever_c_demoted from research: ic=1 ai=1.0]
- alphaXiv
- arXiv
- CatalyzeX
- DagsHub
- Gotit.pub
- Hugging Face
- medical shape foundation model
- radiologist
- ScienceCast
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →