An Advertisement Selection And Playing System Based On Image Recognition

Used in public places advertising playback display can be seen everywhere, but it is always rolling broadcast some fixed advertising information, whether there is user to actually see the advertising, whether the advertising is the needs of users, there is no a good evaluation method. The thesis give a method based on technology. The face image of the user can be recognized gender and age, this will carry out advertising content selection of video playback system. Meanwhile, it is possible to calculate the user facing camera time, then we can estimate the effective time of the user viewing ads, the actual effect of ads will be estimate more accurately.Currently, many with intelligent devices using the Android platform, so the system based on the Android platform to achieve, face recognition technology can be used to infer the user’s favorite. After recognizing user’s age and gender, according to different age and gender automatic selection ads which maybe the user will interested in to play. The system is based on the openCV library, using Haar features for face detection, using the conclusion based on t test and F test, which described the relationship between face image and user’s gender and age to recognize the face image. Through the form of questionnaire survey identified the users with different genders and different ages are interested in advertising content type. This system is combined with HTML5’s new features and native+web Android development mode. The experimental results show that the application of face image gender and age recognition rate is high, the system can select the proper advertising content and push to the user smoothly and the user experience is good.