INFORM-CT uses an LLM planner to create scripts executed by VLMs and segmentation models to automatically detect, classify, and report incidental findings in abdominal CT scans, outperforming pure VLM methods on a benchmark for three organs.
In: European Conference on Computer Vision
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
INFORM-CT: INtegrating LLMs and VLMs FOR Incidental Findings Management in Abdominal CT
INFORM-CT uses an LLM planner to create scripts executed by VLMs and segmentation models to automatically detect, classify, and report incidental findings in abdominal CT scans, outperforming pure VLM methods on a benchmark for three organs.