Name
Title | ||
|---|---|---|
Improving real-time CNN-based pupil detection through domain-specific data augmentation |
Abstract | ||
|---|---|---|
Deep learning is a promising technique for real-world pupil detection. However, the small amount of available accurately-annotated data poses a challenge when training such networks. Here, we utilize non-challenging eye videos where algorithmic approaches perform virtually without errors to automatically generate a foundational data set containing subpixel pupil annotations. Then, we propose multiple domain-specific data augmentation methods to create unique training sets containing controlled distributions of pupil-detection challenges. The feasibility, convenience, and advantage of this approach is demonstrated by training a CNN with these datasets. The resulting network outperformed current methods in multiple publicly-available, realistic, and challenging datasets, despite being trained solely with the augmented eye images. This network also exhibited better generalization w.r.t. the latest state-of-the-art CNN: Whereas on datasets similar to training data, the nets displayed similar performance, on datasets unseen to both networks, ours outperformed the state-of-the-art by ≈27% in terms of detection rate.
|
Year | DOI | Venue |
|---|---|---|
2019 | 10.1145/3314111.3319914 | Proceedings of the 11th ACM Symposium on Eye Tracking Research & Applications |
Keywords | Field | DocType |
data augmentation, deep learning, pupil detection | Computer vision,Computer science,Pupil,Artificial intelligence | Conference |
ISBN | Citations | PageRank |
978-1-4503-6709-7 | 1 | 0.39 |
References | Authors | |
0 | 5 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Shaharam Eivazi | 1 | 1 | 0.39 |
| Thiago Santini | 2 | 3 | 0.76 |
| Alireza Keshavarzi | 3 | 1 | 0.39 |
| Thomas C. Kübler | 4 | 124 | 12.57 |
| Andrea Mazzei | 5 | 1 | 0.39 |