Journal of Multi Disciplinary Engineering Technologies, Vol. 14, Issue 2

Paper Title: Augmented Reality Model for the Virtualisation of the Mask

Authors:Jyoti Arora, Mayank Grover, Kartik Aggarwal

Corresponding Author Email: jyotiarora@msit.in

Abstract: Virtual try on systems for fitting new in shop products have attracted increasing research attention. When the world is in a phase where a virus can spread just through physical contact, automation comes into effect. In this work, we propose Augmented Reality Model for the Virtualisation of the Mask (ARMVM), the image based try-on system for fitting a mask on a person’s face. The human face detection seems to be a difficult problem in computer vision due to its high degree of variability which is why the human face can be called a dynamic object. A desirable model will transform the target mask into the most fitting shape. This transformation is done using the geometric augmentation of images. An optimized method of face detection is performed using the deep convolutional neural network along with the extraction of facial features which further help to place the mask accordingly. ARMVM is suitable for a wide range of mask images.

PDF Link