Dimensions overall score 9.0

GitHub Code Pulse

No public code linked for this paper yet.

• AbSteering, an abnormality-centric framework, is introduced to steer VideoLMs toward precise HRCT report generation.method 95%
• AbSteering incorporates an abnormality-centric Chain-of-Thought scheme to enforce abnormality reasoning.method 90%
• AbSteering utilizes a Direct Preference Optimization objective with clinically confusable abnormalities as hard negatives to improve fine-grained discrimination.method 90%
• General-purpose VideoLMs exhibit strong transferability to high-volume medical imaging when guided by the AbSteering paradigm.result 85%
• AbSteering outperforms state-of-the-art domain-specific CT foundation models pretrained with large-scale CT data.result 90%
• AbSteering achieves superior detection sensitivity in HRCT report generation.result 85%
• AbSteering mitigates hallucinations in HRCT report generation.result 85%
• The data and model weights for AbSteering are publicly available.technical 95%

Competitive landscape

Competitor map is still being generated for this paper. Enable generation or check back soon.

Startup potential card