🤖 AI Summary
Clinical assessment of periorbital distance relies on subjective manual measurements, while existing automated methods suffer from poor generalizability and high data dependency. Method: We propose an end-to-end anatomy-aware AI pipeline featuring PeriorbitAI—a domain-adaptive segmentation model outperforming SAM—integrated with a lightweight classifier and a multimodal fusion architecture, alongside a novel out-of-distribution (OOD) robustness evaluation framework. Contribution/Results: We establish periorbital distance as an interpretable and robust biomarker for ophthalmic disease classification. Our method achieves 77–80% accuracy on in-distribution (ID) data and 63–68% on OOD data—substantially surpassing CNN baselines (14%). Segmentation error remains within inter-reader variability, setting a new clinical benchmark. The pipeline delivers high accuracy, strong generalization across diverse datasets, and inherent clinical interpretability.
📝 Abstract
Periorbital distances are critical markers for diagnosing and monitoring a range of oculoplastic and craniofacial conditions. Manual measurement, however, is subjective and prone to intergrader variability. Automated methods have been developed but remain limited by standardized imaging requirements, small datasets, and a narrow focus on individual measurements. We developed a segmentation pipeline trained on a domain-specific dataset of healthy eyes and compared its performance against the Segment Anything Model (SAM) and the prior benchmark, PeriorbitAI. Segmentation accuracy was evaluated across multiple disease classes and imaging conditions. We further investigated the use of predicted periorbital distances as features for disease classification under in-distribution (ID) and out-of-distribution (OOD) settings, comparing shallow classifiers, CNNs, and fusion models. Our segmentation model achieved state-of-the-art accuracy across all datasets, with error rates within intergrader variability and superior performance relative to SAM and PeriorbitAI. In classification tasks, models trained on periorbital distances matched CNN performance on ID data (77--78% accuracy) and substantially outperformed CNNs under OOD conditions (63--68% accuracy vs. 14%). Fusion models achieved the highest ID accuracy (80%) but were sensitive to degraded CNN features under OOD shifts. Segmentation-derived periorbital distances provide robust, explainable features for disease classification and generalize better under domain shift than CNN image classifiers. These results establish a new benchmark for periorbital distance prediction and highlight the potential of anatomy-based AI pipelines for real-world deployment in oculoplastic and craniofacial care.