Human-AI collaboration boosts medical image segmentation accuracy

Researchers have developed Hi-Seg, a framework that enhances the Segment Anything Model (SAM) for pulmonary nodule segmentation in medical imaging. This human-in-the-loop system allows annotators, including non-medical personnel, to iteratively refine AI-generated masks, leading to improved accuracy and reduced annotation time. In a large-scale study using CT scans from over a thousand patients, Hi-Seg achieved a mean Dice score of nearly 85%, significantly outperforming existing deep learning models and SAM variants. The findings suggest that this collaborative approach can streamline clinical workflows and enable the safe integration of foundation models into medical practice. AI

IMPACT Enhances AI's utility in medical diagnostics by improving segmentation accuracy and reducing clinician workload.

RANK_REASON Academic paper detailing a new method for medical image segmentation. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• computed tomography
• Dice Score
• Hi-Seg
• SAM
• Segment Anything Model

• paper
• product
• other

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