RIT-Eyes: Rendering of near-eye images for eye-tracking applications
This work addresses the problem of reducing manual annotation effort for eye-tracking researchers, though it is incremental by improving upon existing rendering methods.
The authors tackled the need for large annotated datasets for training deep neural networks in video-based eye tracking by developing a synthetic eye image generation platform with enhanced features like an active deformable iris and blinks. They demonstrated its utility by rendering images matching gaze distributions from NVGaze and OpenEDS datasets, achieving competitive performance with semantic segmentation models trained on synthetic data.
Deep neural networks for video-based eye tracking have demonstrated resilience to noisy environments, stray reflections, and low resolution. However, to train these networks, a large number of manually annotated images are required. To alleviate the cumbersome process of manual labeling, computer graphics rendering is employed to automatically generate a large corpus of annotated eye images under various conditions. In this work, we introduce a synthetic eye image generation platform that improves upon previous work by adding features such as an active deformable iris, an aspherical cornea, retinal retro-reflection, gaze-coordinated eye-lid deformations, and blinks. To demonstrate the utility of our platform, we render images reflecting the represented gaze distributions inherent in two publicly available datasets, NVGaze and OpenEDS. We also report on the performance of two semantic segmentation architectures (SegNet and RITnet) trained on rendered images and tested on the original datasets.