An Efficient Robust Eye Localization by Learning the Convolution Distribution Using Eye Template

Figure 6

The structure of the two-level cascade enhancement.