Real-time mask detection in multimedia environment

In many fields, such as financial payment field, some people will wear a mask to avoid face recognition. This is a big problem in the field of online financial payments because it will prevent the development of it. So in this paper, we propose a method of detecting whether a person in an image wears a mask. First, we choose 90 images about different people and we extract their eigenvalues. Then, we employ GMM (Gaussian mixture model) to model these eigenvalues. And we output a probability dense. Finally, we set a threshold to distinguish face and non-face. The experiment verifies the effectiveness of our method.