Research on Responsive Information Interaction Design Based on Facial Recognition Technology

Abstract

The recognition and tracking down moving figures have always been a hot research topic in the field of computer vision. With the development of artificial intelligence in recent years, the application of deep neural networks involves various fields. Compared with the previous solutions, some problems even have better solutions based on deep neural networks. This article mainly discusses targeting people based on face recognition and realizing movement tracking down the target people. The research content of this article can be divided into two aspects. The first aspect is the identification of the target person. Since the overall characteristics of the target individual are not recognizable, and the overall recognition easily leads to errors, this article uses the method of identifying the target person based on the face. This article introduces the application of deep neural networks of facial recognition, and finally selects the SeetaFace faces recognition system that does not depend on any third-party library functions and is open source. The second aspect are the tracking of sports figures. When the algorithm recognizes that the similarity between the image and the target face reaches a certain level, it will automatically switch to tracking the overall torso of the moving person. In the tracking part, a variety of algorithms are compared and compared with experimental data. Experiments have verified that the convergence of face recognition and tracking algorithms is feasible, and this solution not only prevents the interference and occlusion of similar colored objects of the tracking process, but also ensures the stability of the tracking process and is robust., Accuracy and real-time performance have also achieved good results.