Neural Radiance Field (NeRF)

To implement the Neural Radiance Field (NeRF) to synthesize novel views of complex scenes by optimizing an underlying continuous volumetric scene function using a sparse set of input views. NeRF takes input as a single continuous 5D coordinate spatial location (x, y, z) and viewing direction (θ, ϕ) and provides the output as the volume density and emitted radiance which depends on the view.

Dataset

Download the lego data for NeRF from the original author’s link here.

Implementation

To train the NeRF model:

python3 Train.py

Path where the checkpoints will be saved: './Checkpoint/'

To test the NeRF model:

python3 Test.py

Result

References

1. https://rbe549.github.io/spring2023/proj/p2/
2. https://arxiv.org/abs/2003.08934
3. https://pyimagesearch.com/2021/11/17/computer-graphics-and-deep-learning-with-nerf-using-tensorflow-and-keras-part-2/
4. https://colab.research.google.com/github/keras-team/keras-io/blob/master/examples/vision/ipynb/nerf.ipynb