π€ AI Summary
This study addresses the challenge of real-time, accurate identification and anatomical localization of electroencephalography (EEG) electrodes in clinical bedside settings. The authors propose a two-stage visual system: first, a lightweight YOLOv10n model enables real-time electrode detection; second, detected electrodes are mapped to standardized 10β20 system nomenclature via facial landmarks for automatic anatomical validation. This work presents the first approach to achieve real-time EEG electrode detection and 10β20 system registration using only a standard RGB camera, introducing a geometric augmentation strategy to enhance robustness to head rotation and variability in working distance. Evaluated across five subjects, the system achieves a leave-one-subject-out mAP@.5 of 0.94β―Β±β―0.07 and, under leave-one-cap validation, yields mAP@.5 of 0.97 for medium/large caps (0.72 for small caps), operating at 19 FPS on a standard CPU.
π Abstract
We present a two-stage vision system that detects EEG cap electrodes in a live webcam stream and validates their anatomical placement in real time. A single-class YOLO detector localises electrodes; a geometric stage assigns each detection to a named 10-20 role from facial landmarks. Evaluating under subject-disjoint leave-one-subject-out (LOSO) cross-validation across five subjects wearing the clinically-validated Small/Medium/Large caps, the detector attains mAP@.5 = 0.94 +/- 0.07 across five held-out folds (0.96 pooled). A dedicated leave-one-cap-out axis, holding out every frame of a cap regardless of subject, leaves Medium and Large mAP@.5 within 0.01 of LOSO (0.97, 0.97) while Small drops to 0.72 +/- 0.28, a gap confounded with subject familiarity rather than cap style. Geometric augmentation (rotation, perspective, mixup) improves in-plane-roll robustness and temporal-electrode recall at no inference cost, and a landmark-driven head crop extends the usable distance range, lifting mAP@.5 from 0.23 to 0.45 at 0.6 x apparent scale. A compact mobile-candidate backbone (YOLOv10n) keeps the detector at real-time throughput (19 FPS) on a commodity CPU at 640 px.