Real-Time Anti Spoofing Face Detection with Mask Using CNN

Abstract

As COVID-19 spread the whole way across the world, a significant number of us got mindful of how significant face covers are. Medical services authorities and nearby foundations from one side of the planet to the other are encouraging individuals to wear masks ,as it is the best way to forestall the transmission of the infection. Masks have without a doubt frustrated the facial-acknowledgment industry; the innovation has likewise adjusted. It might sound odd yet wearing a cover does not really prevent a PC from recognizing somebody. We are intending to prepare our model to recognize whether the pictures are genuine or fake one even though individuals are wearing face cover. In this paper, we intend to make a liveness detector equipped for spotting counterfeit faces. To make a liveness detector, we will prepare a deep learning neural network fit for recognizing genuine versus counterfeit appearances. It deals with two correlative spaces: RGB space and multi-scale Retinex (MSR) space. The RGB space contains the point-by-point facial surfaces, yet it is sensitive to illumination whereas the MSR pictures can adequately catch the high recurrence data, which is discriminative for face recognition.