Gaze Tracking Using an Unmodified Web Camera and Convolutional Neural Network

Gaze estimation plays a significant role in understating human behavior and human–computer interaction. Currently, there are many methods accessible for gaze estimation. However, most approaches need additional hardware data acquisition which adds an extra cost to tracking. The classic tracking usually require systematic prior knowledge or expertise practical operations. Moreover, they fundamentally based on the characteristics of eye region, utilizing infrared light iris glint track point. It requires high-quality images with particular environmental conditions another source. Recent studies appearance-based have demonstrated capability neural networks, especially convolutional networks (CNN), decode information present achieved significantly simplified In this paper, method that utilizes CNN can be applied various platforms without is presented. An easy fast collection used collecting face eyes from unmodified desktop camera. proposed registered good results; it proves possible predict reasonable accuracy any tools.