The objective of this project was to generate appropriate context of a TV viewer’s interest so that targeted content which interests him/her could be the delivered. The underlying hypothesis is that the content which interests the viewer is the one towards which his/her gaze is fixated the most. For this we implemented an eye gaze tracking system of the TV viewer to track what content interests him/her the most. The algorithm for this task performs iris detection by the use of state-of-the-art Computer Vision and Machine Learning techniques and infers the gaze direction from there.
Further, an Optical Character Recognition system was developed to recognize text that may be present in the Region of Interest (ROI) of the viewer.