Dynamic Asset Replacement in 3D Point Clouds for Immersive VR Environments

Authors: Vaibhav B. Sonawane, Pratik K. Suryawanshi, Prasad R. Sathe, Gokul B. Jadhav, Zainab S. Khanand, Prof. M. V. Kumbharde

Country: India

Full-text Research PDF File:   View   |   Download

Abstract: This research investigates the automated replacement of objects in 3D point clouds with predefined 3D assets to enhance realism in immersive virtual reality (VR) environments. The study addresses challenges such as point cloud data optimization, asset alignment, and interaction fidelity. The replacement process is performed using two approaches: (1) Manual asset replacement utilizing game engines and 3D modeling tools and (2) AI-driven replacement* using Generative Adversarial Networks (GANs) for automatic asset generation. By integrating deep learning-based segmentation and GAN-based asset generation, we improve spatial accuracy and realism in VR applications. The methodology includes preprocessing point cloud data, implementing segmentation techniques, refining asset selection, and optimizing VR integration. Additionally, GAN-generated assets undergo high-performance rendering using dedicated GPUs, ensuring optimal real-time visualization. Results demonstrate significant enhancements in geometric precision, visual fidelity, and performance efficiency, making this approach valuable for applications in gaming, simulation, education, and design.

Keywords:

Paper Id: 232490

Published On: 2025-05-15

Published In: Volume 13, Issue 3, May-June 2025

Share this