8000 GitHub - SerkanDemirci/nerf: Code release for NeRF (Neural Radiance Fields)
[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to content
forked from bmild/nerf

Code release for NeRF (Neural Radiance Fields)

License

Notifications You must be signed in to change notification settings

SerkanDemirci/nerf

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

49 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Improving NeRF: Neural Radiance Fields

Bilkent University

CS 555 - Introduction to Computer Vision Course Project

Spring 2021

By Serkan Demirci

This project contains NeRF implementation (forked), Experimented 3 methods to improve the NeRF model.

Original NeRF project page: http://tancik.com/nerf

My implementation of three methods to improve NeRF:

Methods Implemented

Grid-NeRF Residual

fourier architecture

Grid Embeddings

fourier architecture

Fourier Nerf

fourier architecture

Setup

Python 3 dependencies:

  • Tensorflow 2
  • matplotlib
  • numpy
  • imageio
  • configargparse
  • Tensorflow Addons

Running the code

After installing the dependencies, run the following to download the dataset,

bash download_example_data.sh

Full dataset can be found in: https://drive.google.com/drive/folders/128yBriW1IG_3NJ5Rp7APSTZsJqdJdfc1

To train the default NeRF architecture run,

python run_nerf.py --config project_configs/fern_nerf_config.txt

For the three methods implemented in the project, change the --config argument with the following:

  • project_configs/fern_grid_res_config.txt for Grid-NeRF residual method,
  • project_configs/fern_embeddings_config.txt for Grid Embeddings method,
  • project_configs/fern_fourier_config.txt for Fourier NeRF

You can use tensorboard to view the training progress,

tensorboard --logdir=logs/

About

Code release for NeRF (Neural Radiance Fields)

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Jupyter Notebook 98.3%
  • Python 1.7%
0