Analyzing User Interest Detection in Web sites: A Bimodal Approach

In the world of online navigation, understanding user interaction is crucial. We introduce a unique bimodal model for seeing and analyzing user interest in web adverts in this work. Our innovative method employs real-time camera data, combining face expression analysis and gaze tracking. The presented system utilizes Convolutional Neural Networks (CNNs) to understand more about user emotions while interacting with advertisements. Furthermore, we have built a gaze tracking algorithm that catches and saves gaze coordinates, as well as associated facial emotions, in a CSV file. Surprisingly, our study indicates that the upper center of the homepage receives the most visual attention, while the bottom right quadrant receives the least. With respect to Euclidean distance, the greatest distance measured was 21.4, resulting in the highest path similarity of 42.9%. Furthermore, about 18% has some common screen starting points and 22% terminate at the same screen endpoint. These findings provide an important foundation for improving advertising strategies and optimizing user experience in the changing world of web navigation.