Paper Key : IRJ************565
Author: B. Anjali
Date Published: 02 Feb 2025
Abstract
Easy on the budget is a hallmark of the second-stage training of the AI-based Virtual Garment Trial Room; this unique system uses the latest set of AR and machine-learning-based processes. The body detection very accurately is done using highly efficient Haar cascade algorithms based on OpenCV, while the alignment of the garment uses state-of-the-art convolutional neural networks (CNNs)-creating a very involving and lifelike trial experience through standard webcams. The system favors a significant number of software solutions over a few limited hardware-heavy options like the Kinect on the basis of price and scale. Huge future advancements like Pose Alignment Networks (PAN) and Texture Refinement Networks (TRN) will greatly increase accuracy and realism. The system can greatly reduce the difference between online and offline shopping experiences, bringing the potential of greatly enhancing customer experience and reduce returns. Keywords: OpenCV, machine learning, AR, Haar cascade, Dlib, CNN, PAN, TRN, Kinect.
DOI Requested