Advertisement viewability: A step towards computational advertisement

日期 : Sep 2021 - 现在

<div><p>Advertisements in the malls are running over the loop without effectively observing people in the surroundings, this leads to companies losing millions of rupees because such advertisements do not reach targeted audience effectively and people in the surrounding most of the time do not show interest in such ads. It means companies are not able to attract enough customers towards their brand. Now, to achieve increase viewability and interest of people who visit malls, our Artificial Intelligence based system will detect the people in its surrounding, classify them based on gender and age groups and will display the ads for the gender and age group that is in majority in its surrounding. Our system will periodically monitor surroundings and change ads according to surrounding. We have Trained our models for prediction of gender and age using Haar Cascade for facial features extraction along with Tensor flow for classification of given video input based on gender and age in real time. Trained models are used in the code that enable the correct prediction of gender and age Our system predicts the age with 60 to 65% of the accuracy and predict gender with 80 to 85% of the accuracy, and for each loop of detection an ad display for 5 to 10 seconds depending on how the audience change in the surrounding. Our system will increase the viewability of ads using Artificial Intelligence by effectively detecting the surrounding, leading to generate more sales for same amount of money that companies previously spend over the traditional advertisement.</p></div>