Real-time camera-based eye gaze tracking using convolutional neural network: a case study on social media website

Eye gaze tracking plays an important role in various fields including, human computer interaction, virtual and augmented reality and in identifying effective marketing solutions in affective manner. This paper addresses real-time eye gaze estimation problem using low resolution ordinary camera available in almost every desktop environment as opposed to gaze tracking technologies requiring costly equipment and infrared light sources. In this research, a camera based non-invasive technique has been proposed for tracking and recording gaze points. Further, the proposed framework was used to analyze gaze behavior of users on advertisements displayed on social media website. Eye gaze fixations data of 32 participants were recorded, and gaze patterns were plotted using Heat maps. In addition, the gaze driven interface was designed for virtual interaction tasks to assess the performance, and usability of our proposed framework.