使用在合成数据上训练的深度学习精确定位眼部图像中角膜反射（英文版）

英文标题：Precise localization of corneal reflections in eye images using deep learning trained on synthetic data中文摘要：本文提出了使用深度学习方法基于模拟数据准确定位眼睛图像中单个角膜反射点的中心，且性能比目前算法更好，并可解决目前深度学习模型在视线估计中常见的数据可用性问题，可用于提高基于角膜反射的眼动追踪的精度和准确性。英文摘要：We present a deep learning method for accurately localizing the center of asingle corneal reflection (CR) in an eye image. Unlike previous approaches, weuse a convolutional neural network (CNN) that was trained solely usingsimulated data. Using only simulated data has the benef