Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{172445,
        author = {Akshatha A and Gagana M V and Soumya A},
        title = {Generating 3D objects from 2D images using various NeRF models},
        journal = {International Journal of Innovative Research in Technology},
        year = {2025},
        volume = {11},
        number = {8},
        pages = {3203-3208},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=172445},
        abstract = {This paper introduces a comprehensive framework for enhancing NeRF-based 3D scene modeling by combining advanced deep learning techniques and optimization strategies. The framework supports experimentation with various NeRF backbones, including vanilla, grid-based, and Taichi-based networks, and incorporates diverse raymarching techniques and guidance models like Stable Diffusion, DeepFloyd IF, and Zero123. Key innovations include progressive view expansion, mesh decimation, and improved loss functions for better 3D mesh fidelity and visual quality. The robust training pipeline features mixed precision, adaptive ray marching, and camera pose jittering, facilitating effective handling of complex scenes. Our results show significant improvements in 3D reconstruction quality and efficiency, validated through extensive testing. This framework offers a versatile tool for NeRF technology, with future work aimed at further refinement and real-world application exploration.},
        keywords = {3D Scene Reconstruction, Neural Radiance Fields (NeRF), Progressive View Expansion, Raymarching Techniques, Guidance Models, Stable Diffusion, Zero123Mixed Precision Training},
        month = {January},
        }