New CT scan segmentation method uses weak supervision from text reports

Researchers have developed a novel method for segmenting CT scans using weak supervision derived from textual reports. This approach combines voxel-level supervision with slice-level classification loss extracted from scan-report pairs. By finetuning the SAM3 segmentation model, the technique significantly improves segmentation accuracy, showing a relative gain of up to 22% when using fewer fully labeled volumes. AI

IMPACT This research could lead to more efficient and scalable medical image analysis by reducing reliance on fully labeled datasets.

RANK_REASON The cluster contains an academic paper detailing a new method for CT volume segmentation. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• alphaXiv
• arXiv
• CatalyzeX
• DagsHub
• Gotit.pub
• Hugging Face
• ReXGroundingCT
• SAM3
• Sanjay Subramanian
• ScienceCast

• paper
• model release

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