A structured multi-stage LLM prompt can classify medical segmentation quality zero-shot, with accuracy comparable to trained vision models on a small test set.
Facial video-based non- contact stress recognition utilizing multi-task learning with peak attention
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
eess.IV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Beyond Pixel Agreement: Large Language Models as Clinical Guardrails for Reliable Medical Image Segmentation
A structured multi-stage LLM prompt can classify medical segmentation quality zero-shot, with accuracy comparable to trained vision models on a small test set.