Paper Title
FACE MASK REMOVAL USING GAN MODEL

Abstract
Abstract - This paper is about producing an algorithm capable of detecting face masks, removing them and re-building what is underneath dynamically using a combination of machine learning and mask detection algorithms. To complete this task, an ROI algorithm was built to detect masks of any shape and colour. A pre-trained model created from generative adversarial networks (GAN) was then used to construct the face from scratch once the pixels for the front were removed. The dataset used to train this pre-trained model was the CelebA database, containing 200,000 images of celebrities. Keywords - GAN, ROI, ML, Model, Algorithm