Structured Visual Evidence Decomposition for Evidence-Grounded Multimodal Screening of Obstructive Sleep Apnea-Hypopnea Syndrome
Researchers have developed EviOSAHS, a novel framework designed to improve the screening accuracy for obstructive sleep apnea-hypopnea syndrome (OSAHS). This system decomposes visual evidence from facial images into structured 'evidence cards' that detail anatomical information and risk factors. These cards are then combined with clinical data for a final adjudication by a large language model, achieving an 88.47% accuracy rate in a recent study. AI
IMPACT This framework offers a more auditable and sensitive approach to OSAHS screening, potentially serving as a valuable triage assistant in clinical settings.