Segmentation Agreement and AI-Based Feature Extraction of Cutaneous Infrared Images of the Obese Abdomen after Caesarean Section: Results from a Single Training Session

Background: Infrared thermography in women undergoing caesarean section has promise to identify a surgical site infection prodrome characterised by changes cutaneous perfusion with concomitant influences on temperature distribution across the abdomen. This study was designed compare abdominal and wound regions of interest (ROI) feature extraction agreement between two independent users after single training session. Methods: Image analysis performed manually MATLAB each reviewer ‘blind’ results other. ROIs were annotated via pixel-level segmentation creating pixel masks at four time-points during first 30 days surgery. Results: A total 366 matched image pairs (732 abdomen labels total) obtained. Distribution mask using Jacquard similarity co-efficient ranged from 0.35 1. Good (coefficient ≥ 0.7) (for size shape) observed for abdomen, but poor < 0.7). From extraction, cold spots most those who later developed infections. Conclusions: Reviewer performance, respect input (image) data stage algorithm development, reveals lack correspondence (agreement) ROI indicating need further work refine characteristics output (masks) before an unsupervised works effectively learn patterns features wound.