Saliency Detection in Video Advertisements using Gaze Analysis

Detection of salient features in a video has a wide range of applications. Through gaze tracking, important and interesting features of a video can be highlighted. This data can be used to make the advertisements better and effective. Gaze detection gives an estimation of the points that were under more focus by the viewers. Greater the sample space of the data, better the generic estimation of important points. This data is further parsed through different algorithms to achieve a better accuracy. These algorithms also test the validity of the data or whether the data points collected are sufficient are not. Each algorithm used manipulates the data in a different way to get maximum coverage of all the possible aspects. Finally, the result obtained is a small region of the frame that captures the maximum attention of the viewer