Awesome

GeometrySticker

GeometrySticker: Enabling Ownership Claim of Recolorized Neural Radiance Fields (ECCV 2024)

Paper: arXiv

Project Page: https://kevinhuangxf.github.io/GeometrySticker/

Clone this repository

git clone --branch main --single-branch https://github.com/kevinhuangxf/GeometrySticker.git

Installation

# create conda environment

conda create -n geosticker python=3.8

# install pytorch dependencies

pip install torch==2.1.0 torchvision==0.16.0 torchaudio==2.1.0 --index-url https://download.pytorch.org/whl/cu121

# install tiny-cuda

pip install git+https://github.com/NVlabs/tiny-cuda-nn/#subdirectory=bindings/torch

# install torch-scatter

pip install torch-scatter -f https://data.pyg.org/whl/torch-2.1.0+cu121.html

# install other dependencies

pip install requirements.txt

Dataset

Please download the Blender and LLFF datasets from this link: NeRF datasets.

Run experiments

Run GeometrySticker with NGP-based NeRF.

cd exp/ngp

Train a nerf model.

ROOT_DIR=path/to/Synthetic_NeRF

python train.py \
    --root_dir $ROOT_DIR/Lego \
    --exp_name Lego \
    --num_epochs 30 --batch_size 16384 --lr 2e-2 --eval_lpips

Train GeometrySticker.

python train_geometrysticker.py \
    --root_dir $ROOT_DIR/Lego_geosticker/ \
    --exp_name Lego \
    --lr 1e-4 \
    --num_epochs 5 \
    --weight_path ckpts/nsvf/Lego/epoch=29_slim.ckpt \
    --downsample 0.25

Evaluation

# Evaluating on recoloring

ROOT_DIR=path/to/Synthetic_NeRF

python train_geometrysticker.py \
    --root_dir $ROOT_DIR/Lego_geosticker \
    --exp_name Lego_geosticker \
    --weight_path ckpts/nsvf/Lego_geosticker/epoch=4_slim.ckpt \
    --downsample 0.25 \
    --val_only

Ciatation

@article{huang2024geometrysticker,
  title     = {GeometrySticker: Enabling Ownership Claim of Recolorized Neural Radiance Fields},
  author    = {Xiufeng Huang, Ka Chun Cheung, Simon See, Renjie Wan},
  journal   = {European Conference on Computer Vision (ECCV)},
  year      = {2024},
}