Putting NeRF on a Diet: Semantically Consistent Few-Shot View Synthesis

Last update: Dec 27, 2022

Related tags

Overview

Putting NeRF on a Diet: Semantically Consistent Few-Shot View Synthesis

This repository contains the official implementation of DietNeRF, a system that reconstructs 3D scenes from a few posed photos.

Setup

We use the following folder structure:

dietnerf/
  logs/ (images, videos, checkpoints)
  data/
    nerf_synthetic/
  configs/ (run configuration files)
CLIP/ (Fork of OpenAI's clip repository with a wrapper)

Create conda environment:

conda create -n dietnerf python=3.9
conda activate dietnerf

Set up requirements and our fork of CLIP:

pip install -r requirements.txt
cd CLIP
pip install -e .

wandb login

Experiments on the Realistic Synthetic dataset

Realistic Synthetic experiments are implemented in the ./dietnerf subdirectory.

You need to download datasets from NeRF's Google Drive folder. The dataset was used in the original NeRF paper by Mildenhall et al. For example,

mkdir dietnerf/logs/ dietnerf/data/
cd dietnerf/data
pip install gdown
gdown --id 18JxhpWD-4ZmuFKLzKlAw-w5PpzZxXOcG -O nerf_synthetic.zip
unzip nerf_synthetic.zip
rm -r __MACOSX

Then, shrink images to 400x400:

python dietnerf/scripts/bulk_shrink_images.py "dietnerf/data/nerf_synthetic/*/*/*.png" dietnerf/data/nerf_synthetic_400_rgb/ True

These images are used for FID/KID computation. The dietnerf/run_nerf.py training and evaluation code automatically shrinks images with the --half_res argument.

Each experiment has a config file stored in dietnerf/configs/. Scripts in dietnerf/scripts/ can be run to train and evaluate models. Run these scripts from ./dietnerf. The scripts assume you are running one script at a time on a server with 8 NVIDIA GPUs.

cd dietnerf
export WANDB_ENTITY=
   
    

# NeRF baselines
sh scripts/run_synthetic_nerf_100v.sh
sh scripts/run_synthetic_nerf_8v.sh
sh scripts/run_synthetic_simplified_nerf_8v.sh

# DietNeRF with 8 observed views
sh scripts/run_synthetic_dietnerf_8v.sh
sh scripts/run_synthetic_dietnerf_ft_8v.sh

# NeRF and DietNeRF with partial observability
sh scripts/run_synthetic_unseen_side_14v.sh

Experiments on the DTU dataset

Coming soon. Our paper also fine-tunes pixelNeRF on DTU scenes for 1-shot view synthesis.

Citation and acknowledgements

If DietNeRF is relevant to your project, please cite our associated paper:

@InProceedings{Jain_2021_ICCV,
    author    = {Jain, Ajay and Tancik, Matthew and Abbeel, Pieter},
    title     = {Putting NeRF on a Diet: Semantically Consistent Few-Shot View Synthesis},
    booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)},
    month     = {October},
    year      = {2021},
    pages     = {5885-5894}
}

This code is based on Yen-Chen Lin's PyTorch implementation of NeRF and the official pixelNeRF code.

Putting NeRF on a Diet: Semantically Consistent Few-Shot View Synthesis

Related tags

Overview

Putting NeRF on a Diet: Semantically Consistent Few-Shot View Synthesis

Setup

Experiments on the Realistic Synthetic dataset

Experiments on the DTU dataset

Citation and acknowledgements

Owner

Ajay Jain

Source code for the paper "PLOME: Pre-training with Misspelled Knowledge for Chinese Spelling Correction" in ACL2021

PyMove is a Python library to simplify queries and visualization of trajectories and other spatial-temporal data

Implementation of EMNLP 2017 Paper "Natural Language Does Not Emerge 'Naturally' in Multi-Agent Dialog" using PyTorch and ParlAI

Using contrastive learning and OpenAI's CLIP to find good embeddings for images with lossy transformations

A blender add-on that automatically re-aligns wrong axis objects.

The official implementation of A Unified Game-Theoretic Interpretation of Adversarial Robustness.

Keras implementation of PersonLab for Multi-Person Pose Estimation and Instance Segmentation.

Official implementation of the RAVE model: a Realtime Audio Variational autoEncoder

ObjectDetNet is an easy, flexible, open-source object detection framework

Interactive Terraform visualization. State and configuration explorer.

Code for the paper: Audio-Visual Scene Analysis with Self-Supervised Multisensory Features

Bidimensional Leaderboards: Generate and Evaluate Language Hand in Hand

Source code for TACL paper "KEPLER: A Unified Model for Knowledge Embedding and Pre-trained Language Representation".

The official implementation of Variable-Length Piano Infilling (VLI).

最新版本yolov5+deepsort目标检测和追踪，支持5.0版本可训练自己数据集

This is an official repository of CLGo: Learning to Predict 3D Lane Shape and Camera Pose from a Single Image via Geometry Constraints

NanoDet-Plus⚡Super fast and lightweight anchor-free object detection model. 🔥Only 980 KB(int8) / 1.8MB (fp16) and run 97FPS on cellphone🔥

PyTorch implementation of Densely Connected Time Delay Neural Network

这是一个mobilenet-yolov4-lite的库，把yolov4主干网络修改成了mobilenet，修改了Panet的卷积组成，使参数量大幅度缩小。

Deep Learning to Improve Breast Cancer Detection on Screening Mammography