bmild/nerf

★ 10,870Jupyter NotebookAudience · researcherComplexity · 5/5Setup · hard

Mindmap

mindmap
  root((NeRF))
    What it does
      Novel view synthesis
      3D scene from photos
      Neural radiance field
    Inputs
      Multi-angle photos
      Camera positions
    Outputs
      New view renders
      Fly-through video
      3D mesh export
    Tech Stack
      Python
      TensorFlow 1.15
      COLMAP
    Audience
      AI researchers
      Computer vision devs

mindmap root((NeRF)) What it does Novel view synthesis 3D scene from photos Neural radiance field Inputs Multi-angle photos Camera positions Outputs New view renders Fly-through video 3D mesh export Tech Stack Python TensorFlow 1.15 COLMAP Audience AI researchers Computer vision devs

Click or tap to explore — scroll the page freely

Things people build with this

USE CASE 1

Reproduce the original NeRF paper results on the included fern or Lego scenes to validate your GPU environment before experimenting.

USE CASE 2

Train a NeRF model on your own set of photos to generate a smooth fly-through video of a real-world scene from novel angles.

USE CASE 3

Extract a 3D mesh from a trained NeRF model using the included notebook to use in a 3D application or game engine.

USE CASE 4

Use this reference implementation as a baseline to benchmark a new neural rendering method against the original paper.

Tech stack

PythonTensorFlowJupyter Notebookconda

Getting it running

Difficulty · hard Time to first run · 1day+

Requires a CUDA GPU, TensorFlow 1.15, a conda environment, and separate tools (LLFF and COLMAP) to process custom photos, training takes hours to days.

No explicit license stated in the explanation.

In plain English

NeRF (Neural Radiance Fields) is research code released alongside an academic paper from UC Berkeley and Google Research, published at the ECCV 2020 conference where it received a Best Paper Honorable Mention. The project introduces a method for generating new views of a scene from photos taken from different angles, using a neural network trained on those photos. The core idea is that you give the system a set of photos of the same physical scene taken from various positions. It trains a small neural network (about 5MB in size) that learns to represent that scene in 3D. Once trained, the network can generate a realistic photo from any new camera angle that was never actually photographed. The result is a smooth video that appears to fly through or around the scene. Training takes anywhere from a few hours to a day or two depending on the image resolution, and requires a single GPU. Rendering a new image from the trained model can take anywhere from under a second to about 30 seconds, also depending on resolution. The code includes example scenes (a fern plant and a Lego model) so you can run a demonstration without your own data. The code is implemented using TensorFlow 1.15 and runs in Python 3. Setup is handled through a conda environment file. If you want to use your own photos rather than the included examples, you need to compute camera positions using additional tools (LLFF and COLMAP) that are separate projects referenced in the README. The repository also includes a notebook for extracting 3D geometry as a mesh from a trained model. This is the original reference implementation from the paper's authors. It is primarily useful for researchers or technically proficient users who want to reproduce the paper's results or experiment with the method. Casual use requires comfort with Python environments, the command line, and GPU hardware.

Copy-paste prompts

Prompt 1

How do I set up the NeRF conda environment and train the model on the included fern scene to render a novel-view video?

Prompt 2

Walk me through preparing my own photos for NeRF using LLFF and COLMAP to compute camera positions before training.

Prompt 3

How long does training NeRF take on a single GPU, and what image resolution gives a good balance between quality and training time?

Prompt 4

Show me how to use the NeRF Jupyter notebook to extract a 3D mesh from a trained model so I can import it into Blender.

Prompt 5

What are the main differences between the original NeRF implementation and modern derivatives like Instant-NGP, and when should I use the original?

Open on GitHub → Explain another repo

← bmild on gitmyhub — every repo by this author, as a profile.

Verify against the repo before relying on details.