An Approach of Face Detection with Cap and Mask

Abstract

Face detection is a technology that is broadly implemented and studied on deep learning with hugely improved face detection system indicators. Extensive research has been done to detect faces under illumination changes, and the resulting degenerative image problems are largely ignored. Maximum existing face detection methods focus on overcoming the problem of facial obstruction due to sunglasses. As far we know, no research has been done on obstruction due to cap and mask. Yet this problem should emphasize because it is known that bank criminals adopt this process to hide their faces. Our system offers purposes for solving this problem. Face detection with cap and mask applying Deep Learning is a highly challenging job. Using the TensorFlow Object Detection select for dataset training in the deep learning model. Specifically, we are going to use a pre-trained model called SSD-MobileNet. This method applies a camera, which captures images. First, Prepare the image labels using Labelimg. Then, Transfer learning using SSD MobileNet. Finally, Detection using a webcam and OpenCV and display the text and percentage. The implementation of the process is by using OpenCVPython. The technique applies to various libraries.