๐Ÿ˜ฎThe official implementation of "CoNeRF: Controllable Neural Radiance Fields" ๐Ÿ˜ฎ

Overview

CoNeRF: Controllable Neural Radiance Fields

arXiv MIT license Website Datasets

This is the official implementation for "CoNeRF: Controllable Neural Radiance Fields"

The codebase is based on HyperNeRF implemente in JAX, building on JaxNeRF.

Setup

The code can be run under any environment with Python 3.8 and above. (It may run with lower versions, but we have not tested it).

We recommend using Miniconda and setting up an environment:

conda create --name conerf python=3.8

Next, install the required packages:

pip install -r requirements.txt

Install the appropriate JAX distribution for your environment by following the instructions here. For example:

# For CUDA version 11.1
pip install --upgrade "jax[cuda111]" -f https://storage.googleapis.com/jax-releases/jax_releases.html

Dataset

Basic structure

The dataset uses the same format as Nerfies for the image extraction and camera estimation.

For annotations, we create an additional file annotations.yml consisting of attribute values and their corresponding frames, and a folder with [frame_id].json files (only annotated frames are required to have a corresponding .json file) where each *.json file is a segmentation mask created with LabelMe. In summary, each dataset has to have the following structure:

<dataset>
    โ”œโ”€โ”€ annotations
    โ”‚   โ””โ”€โ”€ ${item_id}.json
    โ”œโ”€โ”€ annotations.yml
    โ”œโ”€โ”€ camera
    โ”‚   โ””โ”€โ”€ ${item_id}.json
    โ”œโ”€โ”€ camera-paths
    โ”œโ”€โ”€ colmap
    โ”œโ”€โ”€ rgb
    โ”‚   โ”œโ”€โ”€ ${scale}x
    โ”‚   โ””โ”€โ”€ โ””โ”€โ”€ ${item_id}.png
    โ”œโ”€โ”€ metadata.json
    โ”œโ”€โ”€ dataset.json
    โ”œโ”€โ”€ scene.json
    โ””โ”€โ”€ mapping.yml

The mapping.yml file can be created manually and serves to map class indices to class names which were created with LabelMe. It has the following format:

<index-from-0>: <class-name>

for example:

0: left eye
1: right eye

The annotations.yml can be created manually as well (though we encourage using the provided notebook for this task) and has the following format:

- class: <id>
  frame: <number>
  value: <attribute-value> # between -1 and 1

for example:

- class: 0 # corresponding to left eye
  frame: 128
  value: -1
- class: 1 # corresponding to right eye
  frame: 147
  value: 1
- class: 2 # corresponding to mouth
  frame: 147
  value: -1 

Principles of annotating the data

  • Our framework works well with just a bunch of annotations (for extreme points as an example). For our main face visualizations, we used just 2 annotations per attribute.
  • We highly recommend annotating these frames that are extremes of possible controllability, for example, fully eye closed will be -1 value and fully open eye will +1 value. Though it is not necessary to be exact in extremes, the more accurate annotations, the more accurate controllability you can expect
  • Each attribute can be annotated independently, i.e., there is no need to look for frames that have exactly extreme values of all attributes. For example, left eye=-1 and left eye=+1 values can be provided in frames 28 and 47, while right eye=-1 and right eye=+1 can be provided in any other frames.
  • Masks should be quite rough oversized, it is generally better to have bigger than smaller annotations.
  • The general annotation pipeline looks like this:
  1. Find set of frames that consist of extreme attributions (e.g. closed eye, open eye etc.).
  2. Provide necessary values in for attributes to be controlled in annotations.yml.
  3. Set names for these attributes (necessary for the masking part).
  4. Run LabelMe.
  5. Save annotated frames in annotations/.

Now you can run the training! Also, check out our datasets (52GB of data) to avoid any preprocessing steps on your own.

We tried our best to make our CoNeRF codebase to be general for novel view synthesis validation dataset (conerf/datasets/nerfies.py file) but we mainly focused on the interpolation task. If you have an access to the novel view synthesis rig as used in NeRFies or HyperNeRF, and you find out that something doesn't work, please leave an issue.

Providing value annotations

We extended the basic notebook used in NeRFies and HyperNeRF for processing the data so that you can annotate necessary images with attributes. Please check out notebooks/Capture_Processing.ipynb for more details. The notebook (despite all the files from NeRFies) will also generate <dataset>/annotations.yml and <dataset>/mapping.yml files.

Providing masking annotations

We adapted data loading class to handle annotations from LabelMe (we used its docker version). Example annotation for one of our datasets looks like this:

example-annotation

The program generates *.json files in File->Output Dir which should be located in <dataset>/annotations/ folder.

Training

After preparing a dataset, you can train a Nerfie by running:

export DATASET_PATH=/path/to/dataset
export EXPERIMENT_PATH=/path/to/save/experiment/to
python train.py \
    --base_folder $EXPERIMENT_PATH \
    --gin_bindings="data_dir='$DATASET_PATH'" \
    --gin_configs configs/test_local_attributes.gin

To plot telemetry to Tensorboard and render checkpoints on the fly, also launch an evaluation job by running:

python eval.py \
    --base_folder $EXPERIMENT_PATH \
    --gin_bindings="data_dir='$DATASET_PATH'" \
    --gin_configs configs/test_local_attributes.gin

The two jobs should use a mutually exclusive set of GPUs. This division allows the training job to run without having to stop for evaluation.

Configuration

  • We use Gin for configuration.
  • We provide a couple preset configurations.
  • Please refer to config.py for documentation on what each configuration does.
  • Preset configs:
    • baselines/: All configs that were used to perform quantitative evaluation in the experiments, including baseline methods. The _proj suffix denotes a method that uses a learnable projection.
      • ours.gin: The full CoNeRF architecture with masking.
      • hypernerf_ap[_proj].gin: The axis-aligned plane configuration for HyperNeRF.
      • hypernerf_ds[_proj].gin: The deformable surface configuration for HyperNeRF.
      • nerf_latent[_proj].gin: The configuration for a simple baselines where we concatenate a learnable latent with each coordinate (resembles HyperNeRF AP without the warping field).
      • nerfies[_proj].gin: The configuration for the NeRFies model.
      • nerf.gin: The configuration for the simplest NeRF architecture.
    • full-hd/, hd/ and post/: We repurposed our baselines/ours.gin configuration for training for different resolutions and different sampling parameters that increase the quality of the generated images. Using post/ours.gin required us to use 4x A100 GPU for 2 weeks to make the training converge.

Synthetic dataset

We generated the synthetic dataset using Kubric. You can find the generation script here. After generating the dataset, you can run prepare_kubric_dataset.py to canonicalize its format to the same one that works with CoNeRF. The dataset is already attached in the provided zip file.

Additional scripts

All scripts below are used as the ones for training, they need $EXPERIMENT_PATH and $DATASET_PATH to be specified. They save the results into $EXPERIMENT_PATH.

  • render_changing_attributes.py: Renders each of changing attributes under a fixed camera.
  • render_video.py: Renders changing view under a fixed set of attributes.
  • render_all.py: Renders dynamically changing attributes and the camera parameters.
  • train_lr.py: Estimates parameters of the linear regression. The estimated model maps highly dimensional embedding into controllable attributes.

Additional notes

  • We have used notebooks/Results.ipynb to generate tables/visualizations for the article. While it may not particularily useful for you case, we have left it so you can copy or reuse some of its snippets. It's especially useful because it shows how to extract data from tensorboards.
  • We removed some of notebooks that were available in the HyperNeRF's codebase (ex. for training) but were no longer applicable to CoNeRF. We highly recommend using available scripts. If you have ever managed to adapt HyperNeRF's notebooks, please leave a pull request.

Citing

If you find our work useful, please consider citing:

@inproceedings{kania2022conerf,
  title     = {{CoNeRF: Controllable Neural Radiance Fields}},
  author    = {Kania, Kacper and Yi, Kwang Moo and Kowalski, Marek and Trzci{\'n}ski, Tomasz and Tagliasacchi, Andrea},
  booktitle   = {Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition},
  year      = {2022}
}
Owner
Kacper Kania
PhDing in Neural Human Rendering ... ๐Ÿ‘€
Kacper Kania
PHOTONAI is a high level python API for designing and optimizing machine learning pipelines.

PHOTONAI is a high level python API for designing and optimizing machine learning pipelines. We've created a system in which you can easily select and

Medical Machine Learning Lab - University of Mรผnster 57 Nov 12, 2022
ALBERT-pytorch-implementation - ALBERT pytorch implementation

ALBERT-pytorch-implementation developing... ๋ชจ๋ธ์˜ ๊ฐœ๋…์ดํ•ด๋ฅผ ๋•๊ธฐ ์œ„ํ•œ ๊ตฌํ˜„๋ฌผ๋กœ ํ˜„์žฌ ๋ณ€์ˆ˜๋ช…์„ ์ƒ์„ธํžˆ ์ ์—ˆ๊ณ 

BG Kim 3 Oct 06, 2022
PyTorch implementation of Neural View Synthesis and Matching for Semi-Supervised Few-Shot Learning of 3D Pose

Neural View Synthesis and Matching for Semi-Supervised Few-Shot Learning of 3D Pose Release Notes The official PyTorch implementation of Neural View S

Angtian Wang 20 Oct 09, 2022
Examples of using f2py to get high-speed Fortran integrated with Python easily

f2py Examples Simple examples of using f2py to get high-speed Fortran integrated with Python easily. These examples are also useful to troubleshoot pr

Michael 35 Aug 21, 2022
ROS support for Velodyne 3D LIDARs

Overview Velodyne1 is a collection of ROS2 packages supporting Velodyne high definition 3D LIDARs3. Warning: The master branch normally contains code

ROS device drivers 543 Dec 30, 2022
[ICLR 2022] Contact Points Discovery for Soft-Body Manipulations with Differentiable Physics

CPDeform Code and data for paper Contact Points Discovery for Soft-Body Manipulations with Differentiable Physics at ICLR 2022 (Spotlight). @InProceed

(Lester) Sizhe Li 29 Nov 29, 2022
This is the code repository for the paper "Identification of the Generalized Condorcet Winner in Multi-dueling Bandits" (NeurIPS 2021).

Code Repository for the Paper "Identification of the Generalized Condorcet Winner in Multi-dueling Bandits" (To appear in: Proceedings of NeurIPS20

1 Oct 03, 2022
OpenLT: An open-source project for long-tail classification

OpenLT: An open-source project for long-tail classification Supported Methods for Long-tailed Recognition: Cross-Entropy Loss Focal Loss (ICCV'17) Cla

Ming Li 37 Sep 15, 2022
Implementation of "Debiasing Item-to-Item Recommendations With Small Annotated Datasets" (RecSys '20)

Debiasing Item-to-Item Recommendations With Small Annotated Datasets This is the code for our RecSys '20 paper. Other materials can be found here: Ful

Microsoft 34 Aug 10, 2022
Official MegEngine implementation of CREStereo(CVPR 2022 Oral).

[CVPR 2022] Practical Stereo Matching via Cascaded Recurrent Network with Adaptive Correlation This repository contains MegEngine implementation of ou

MEGVII Research 309 Dec 30, 2022
A modular application for performing anomaly detection in networks

Deep-Learning-Models-for-Network-Annomaly-Detection The modular app consists for mainly three annomaly detection algorithms. The system supports model

Shivam Patel 1 Dec 09, 2021
This repository implements Douzero's interface to IGCA.

douzero-interface-for-ICGA This repository implements Douzero's interface to ICGA. ./douzero: This directory stores Doudizhu AI projects. ./interface:

zhanggenjin 4 Aug 07, 2022
DCA - Official Python implementation of Delaunay Component Analysis algorithm

Delaunay Component Analysis (DCA) Official Python implementation of the Delaunay

Petra Poklukar 9 Sep 06, 2022
Framework for joint representation learning, evaluation through multimodal registration and comparison with image translation based approaches

CoMIR: Contrastive Multimodal Image Representation for Registration Framework ๐Ÿ–ผ Registration of images in different modalities with Deep Learning ๐Ÿค–

Methods for Image Data Analysis - MIDA 55 Dec 09, 2022
Relative Positional Encoding for Transformers with Linear Complexity

Stochastic Positional Encoding (SPE) This is the source code repository for the ICML 2021 paper Relative Positional Encoding for Transformers with Lin

Antoine Liutkus 48 Nov 16, 2022
DeconvNet : Learning Deconvolution Network for Semantic Segmentation

DeconvNet: Learning Deconvolution Network for Semantic Segmentation Created by Hyeonwoo Noh, Seunghoon Hong and Bohyung Han at POSTECH Acknowledgement

Hyeonwoo Noh 325 Oct 20, 2022
Linear image-to-image translation

Linear (Un)supervised Image-to-Image Translation Examples for linear orthogonal transformations in PCA domain, learned without pairing supervision. Tr

Eitan Richardson 40 Aug 31, 2022
PConv-Keras - Unofficial implementation of "Image Inpainting for Irregular Holes Using Partial Convolutions". Try at: www.fixmyphoto.ai

Partial Convolutions for Image Inpainting using Keras Keras implementation of "Image Inpainting for Irregular Holes Using Partial Convolutions", https

Mathias Gruber 871 Jan 05, 2023
Medical Image Segmentation using Squeeze-and-Expansion Transformers

Medical Image Segmentation using Squeeze-and-Expansion Transformers Introduction This repository contains the code of the IJCAI'2021 paper 'Medical Im

askerlee 172 Dec 20, 2022