This is an official repository of the ManifoldNeRF accepted BMVC2023.　　 [arxiv]

Environment

• Python 3.8.20
• CUDA 12.4

Setup

Prerequisites

• Python 3.8.20
• CUDA 12.4

Installation Steps

1. Make conda environment

   conda create -n manifoldnerf python=3.8.20
   conda activate manifoldnerf
   

2. Install PyTorch

   pip install torch==2.4.1 torchvision==0.19.1 torchaudio==2.4.1 --index-url https://download.pytorch.org/whl/cu124
   

3. Install CLIP (Editable Mode)

   cd CLIP
   pip install -e .
   cd ..
   

4. Install Dependencies

   pip install -r requirements.txt
   

5. Download Dataset

   • Download the NeRF synthetic dataset from here
   • Extract and place it in the /data/ directory

Usage

LEGO Scene Example

python manifoldnerf/run_manifold_nerf.py --config config/nerf_synthetic_dataset/mn_blender_chair_8views_ctr.txt --stop_wandb_logging

Please note that you need to edit the dataset path and log path in the config file to match your actual environment.

Citation and acknowledgements

If ManifoldNeRF is relevant to your project, please cite our associated paper:

@inproceedings{Kanaoka_2023_BMVC,
author    = {Daiju Kanaoka and Motoharu Sonogashira and Hakaru Tamukoh and Yasutomo Kawanishi},
title     = {ManifoldNeRF: View-dependent Image Feature Supervision for Few-shot Neural Radiance Fields},
booktitle = {34th British Machine Vision Conference 2023, {BMVC} 2023, Aberdeen, UK, November 20-24, 2023},
publisher = {BMVA},
year      = {2023},
url       = {https://papers.bmvc2023.org/0682.pdf}
}

This code is based on official implementation of DietNeRF