Images of the eye are key in several computer vision problems, such as shape registration and gaze estimation. Recent large-scale supervised methods for these problems require time-consuming data collection and manual annotation, which can be unreliable. We propose synthesizing perfectly labelled photo-realistic training data in a fraction of the time. We used computer graphics techniques to build a collection of dynamic eye-region models from head scan geometry. These were randomly posed to synthesize close-up eye images for a wide range of head poses, gaze directions, and illumination conditions. Finally, we demonstrate the benefits of our synthesized training data (SynthesEyes) by out-performing state-of-the-art methods for eye-shape registration as well as cross-dataset appearance-based gaze estimation in the wild.
The dataset contains 11,382 synthesized close-up images of eyes. There are ten directories, one for each dynamic eye region model in our collection. Each eye image has associated data stored in a pickle file. The directory structure for the dataset is as follows:
SynthesEyes_dataset ├── f01 # data for f01 eye region model │ ├── f01_36_0.1963_-0.7854.png # 120x80px image │ ├── f01_36_0.1963_-0.7854.pkl # associated data for that image │ └── … ├── f02 … # data for f03, f04 … m03, m04 └── m05
The associated data for each image is a dict with keys:
@inproceedings{wood2015_iccv, title = {Rendering of Eyes for Eye-Shape Registration and Gaze Estimation}, author = {Erroll Wood and Tadas Baltrusaitis and Xucong Zhang and Yusuke Sugano and Peter Robinson and Andreas Bulling}, year = {2015}, date = {2015-12-12}, booktitle = {Proc. of the IEEE International Conference on Computer Vision (ICCV 2015)} }
