Event Enhanced Gaussian Splatting(E2GS)

Hiroyuki Deguchi*, Mana Masuda*, Takuya Nakabayshi, Hideo Saito (* indicates equal contribution)

Requirements and Setup

Our code is based on 3D Gaussian Splatting. Please refer to 3D Gaussian Splatting Requirements and Setup.

Data

We use E2NeRF synthetic dataset, so first dowload it. After that, use COLMAP to make the point cloud. As for the synthetic data, we have to generate only a point cloud, so it works even if we input blury image directly to COLMAP. However, if you prefer to use deblurred images, it would be good to refer to EDI model. Data Structere is like below.

|--data
  |--chair
    |--images
      |--r_0.png
      |--r_2.png
      |--...
    |--sparse
      |--0
        |--points3D.bin
    |--events.pt
    |--transform_test.json
    |--transform_train.json
  |--ficus
  |--...
  

As for the realworld data, we have to deblur the blured image using EDI model and estimate 5 camera poses in exposure time per image.

The data I have processed can be downloaded from the following link.

Download the input data

(images_d in each directory is a directory which contains images deblurred by EDI model.)

Running

To run the optimizer, simply use

python train.py -s <path to COLMAP or NeRF Synthetic dataset>

Command Line Arguments for train.py

--source_path / -s

Path to the source directory containing a COLMAP or Synthetic NeRF data set.

--model_path / -m

Path where the trained model should be stored (output/<random> by default).

--images / -i

Alternative subdirectory for COLMAP images (images by default).

--eval

Add this flag to use a MipNeRF360-style training/test split for evaluation.

--resolution / -r

Specifies resolution of the loaded images before training. If provided 1, 2, 4 or 8, uses original, 1/2, 1/4 or 1/8 resolution, respectively. For all other values, rescales the width to the given number while maintaining image aspect. If not set and input image width exceeds 1.6K pixels, inputs are automatically rescaled to this target.

--data_device

Specifies where to put the source image data, cuda by default, recommended to use cpu if training on large/high-resolution dataset, will reduce VRAM consumption, but slightly slow down training. Thanks to HrsPythonix.

--white_background / -w

Add this flag to use white background instead of black (default), e.g., for evaluation of NeRF Synthetic dataset.

--sh_degree

Order of spherical harmonics to be used (no larger than 3). 3 by default.

--convert_SHs_python

Flag to make pipeline compute forward and backward of SHs with PyTorch instead of ours.

--convert_cov3D_python

Flag to make pipeline compute forward and backward of the 3D covariance with PyTorch instead of ours.

--debug

Enables debug mode if you experience erros. If the rasterizer fails, a dump file is created that you may forward to us in an issue so we can take a look.

--debug_from

Debugging is slow. You may specify an iteration (starting from 0) after which the above debugging becomes active.

--iterations

Number of total iterations to train for, 30_000 by default.

--ip

IP to start GUI server on, 127.0.0.1 by default.

--port

Port to use for GUI server, 6009 by default.

--test_iterations

Space-separated iterations at which the training script computes L1 and PSNR over test set, 7000 30000 by default.

--save_iterations

Space-separated iterations at which the training script saves the Gaussian model, 7000 30000 <iterations> by default.

--checkpoint_iterations

Space-separated iterations at which to store a checkpoint for continuing later, saved in the model directory.

--start_checkpoint

Path to a saved checkpoint to continue training from.

--quiet

Flag to omit any text written to standard out pipe.

--feature_lr

Spherical harmonics features learning rate, 0.0025 by default.

--opacity_lr

Opacity learning rate, 0.05 by default.

--scaling_lr

Scaling learning rate, 0.005 by default.

--rotation_lr

Rotation learning rate, 0.001 by default.

--position_lr_max_steps

Number of steps (from 0) where position learning rate goes from initial to final. 30_000 by default.

--position_lr_init

Initial 3D position learning rate, 0.00016 by default.

--position_lr_final

Final 3D position learning rate, 0.0000016 by default.

--position_lr_delay_mult

Position learning rate multiplier (cf. Plenoxels), 0.01 by default.

--densify_from_iter

Iteration where densification starts, 500 by default.

--densify_until_iter

Iteration where densification stops, 15_000 by default.

--densify_grad_threshold

Limit that decides if points should be densified based on 2D position gradient, 0.0002 by default.

--densification_interval

How frequently to densify, 100 (every 100 iterations) by default.

--opacity_reset_interval

How frequently to reset opacity, 3_000 by default.

--lambda_dssim

Influence of SSIM on total loss from 0 to 1, 0.2 by default.

--percent_dense

Percentage of scene extent (0--1) a point must exceed to be forcibly densified, 0.01 by default.

Evaluation

By default, the trained models use all available images in the dataset. To train them while withholding a test set for evaluation, use the --eval flag. This way, you can render training/test sets as follows:

python render.py -m <path to trained model> --eval # Generate input view renderings
python render_novel.py -m <path to trained model> --eval # Generate novel view renderings

If you put the GT clear images, you can compute the metrics error as follows:

python metrics.py -m <path to trained model> # Compute error metrics on renderings

Citation

@inproceedings{E2GS,
title={E2GS:Event Enhanced Gaussian Splatting},
author={Hiroyuki Deguchi and Mana Masuda and Takuya Nakabayashi and Hideo Saito},
booktitle={IEEE International Conference on Image Processing},
year={2024}
}