drealarun / nerflow

[ICCV'21] Neural Radiance Flow for 4D View Synthesis and Video Processing

NeRFlow

[ICCV'21] Neural Radiance Flow for 4D View Synthesis and Video Processing

Datasets

The pouring dataset used for experiments can be download at this URL and the iGibson dataset used in experiments can be downloaded at this URL

Pouring Dataset

Please download and extract each dataset at data/nerf_synthetic/. Please use the following command to train

python run_nerf.py --config=configs/pour_baseline.txt

After running model for 200,000 iterations, move the model to a new folder pour_dataset_flow and then use the following command to train with flow consistency

python run_nerf.py --config=configs/pour_baseline_flow.txt

Gibson Dataset

Please download and extract each dataset at data/nerf_synthetic/. Please use the following command to train the model

python run_nerf.py --config=configs/gibson_baseline.txt

After running model for 200,000 iterations, move the model to a new folder pour_dataset_flow and then use the following command to train with flow consistency

python run_nerf.py --config=configs/gibson_baseline_flow.txt

Citing our Paper

If you find this repo helpful, please consider citing

 @inproceedings{du2021nerflow,
                      author    = {Yilun Du and Yinan Zhang and Hong-Xing Yu 
                                   and Joshua B. Tenenbaum and Jiajun Wu},
                      title     = {Neural Radiance Flow for 4D View Synthesis and Video Processing},
                      year      = {2021},
                      booktitle   = {Proceedings of the IEEE/CVF International Conference
                                     on Computer Vision},
                    }