Low Cost Eye Tracking: The Current Panorama

Figure 1

The common software structure for visible light gaze trackers. The methods start by locating the eyes. To make the estimation more stable, spatiotemporal tracking may be utilized at this step. Later, the location information is used to extract features, fit 2D or 3D models to the eyes, or just extract the eye region image. In the case of model-based methods, the fitted model is used to calculate the gaze geometrically, whereas in the other methods, a mapping function is necessary to calculate the gaze angle or point.