Gaze Estimation Based on Neural Network
Mingyuan Luo, Liu Xi, Wei Huang · 2019
In this study, we selected some state-of-the-art or classic neural network models for recently popular research problem gaze estimation. Experimental analysis based on eyegaze data set compares different models and suggests that neural networks can achieve excellent performance in gaze estimation problem while original Capsule Network provides the best performance and 18-layers ResNet is second-best in all models.