Semi-Supervised Learning for Eye Image Segmentation
This work addresses the labeling bottleneck in eye tracking for applications like human-computer interaction, but it is incremental as it builds on existing semi-supervised methods with domain-specific adaptations.
The paper tackled the problem of eye image segmentation for eye tracking by developing semi-supervised learning frameworks to reduce the need for large labeled datasets, achieving improvements of 0.38% and 0.65% in segmentation performance over a baseline when trained on only 48 labeled images.
Recent advances in appearance-based models have shown improved eye tracking performance in difficult scenarios like occlusion due to eyelashes, eyelids or camera placement, and environmental reflections on the cornea and glasses. The key reason for the improvement is the accurate and robust identification of eye parts (pupil, iris, and sclera regions). The improved accuracy often comes at the cost of labeling an enormous dataset, which is complex and time-consuming. This work presents two semi-supervised learning frameworks to identify eye-parts by taking advantage of unlabeled images where labeled datasets are scarce. With these frameworks, leveraging the domain-specific augmentation and novel spatially varying transformations for image segmentation, we show improved performance on various test cases. For instance, for a model trained on just 48 labeled images, these frameworks achieved an improvement of 0.38% and 0.65% in segmentation performance over the baseline model, which is trained only with the labeled dataset.