Cross-domain evaluation and fine-tuned adaptation of iCatcher+ for Korean infant gaze data

This study evaluates the cross-domain generalisability of three iCatcher+ gaze classification models on a dataset of Korean infants. Applying the pretrained models to the Korean dataset resulted in a substantial decline in accuracy across all models. Performance was systematically modulated by developmental mismatch: accuracy declined as the age distribution of the target dataset diverged from that of the training data. In contrast, experiment-level factors such as stimulus type and temporal progression had limited and model-dependent effects. Fine-tuning the Lookit model on Korean data improved target-domain performance, but at the cost of reduced accuracy on the original source dataset, consistent with catastrophic forgetting. Our findings highlight that deep learning-based gaze classifiers remain sensitive to domain shift. Although fine-tuning can partially mitigate performance loss, robust generalisation may require closer alignment between training and deployment contexts.