Web-based image-Augmented Reality (AR) matching generator

• Main Author: Bill Hundson, David
• Format: Final Year Project / Dissertation / Thesis
• Published: 2023
• Subjects: T Technology (General); TA Engineering (General). Civil engineering (General)
• Online Access: http://eprints.utar.edu.my/5989/1/fyp_IA_2023_BHD.pdf; http://eprints.utar.edu.my/5989/

ABSTRACT This is a project paper on the development for a cross-platform augmented reality application. To achieve its cross-platform nature, the completed system will be developed as a web-based application. In addition to using standard JavaScript, the project will also be incorporating React.js and Node.js to develop the system. In a nutshell, the project aims to deliver a system for video conferencing albeit with enhanced functionalities through exploiting the potentials of AR technology. Multiple video streams of users will be processed in real-time to extract the prominent human subject in a given video scene. The extracted video image of multiple human subjects will then be rendered and displayed within a single video scene. This is the process of augmenting user video images onto a virtual environment. The virtual environment, also referred to as the background element of the rendered output video, will be customizable to a certain degree. This system offers a unique software solution to deliver an AR experience with many creative implications. The completed system will incorporate the Selfie Segmentation Model developed by MediaPipe to perform the segmentation process in order to separate the foreground and background elements of a given video stream. The foreground subject refers to the human subject in a video. The completed system will allow manipulation of the background element. Other than that, the system will also utilize WebRTC technology to facilitate the video communication aspect between users of the system. The successfully developed system will be deployed to the web after which it will be publicly accessible through a given URL.