DE-TensoRF - a data-efficient implementation of TensoRF. This is a course project for CPSC533R: Computer Graphics and Computer Vision.
Proposed three techniques to achieve data-efficiency:
- Obtain unseen views using symmetry
- Semantic conditioning
- Semantic loss
Detailed report can be found here
This work is based on TensoRF. The original code can be found here.
Install environment:
conda create -n TensoRF python=3.8
conda activate TensoRF
pip install torch torchvision
pip install tqdm scikit-image opencv-python configargparse lpips imageio-ffmpeg kornia lpips tensorboard
python train.py --config configs/lego.txt