Deep Illuminator is a data augmentation tool designed for image relighting. It can be used to easily and efficiently generate a wide range of illumination variants of a single image.

Overview

Deep Illuminator

Deep Illuminator is a data augmentation tool designed for image relighting. It can be used to easily and efficiently generate a wide range of illumination variants of a single image. It has been tested with several datasets and models and has been shown to succesfully improve performance. It has a built in visualizer created with Streamlit to preview how the target image can be relit. This tool has an accompanying paper.

Example Augmentations

Usage

The simplest method to use this tool is through Docker Hub:

docker pull kartvel/deep-illuminator

Visualizer

Once you have the Deep Illuminator image run the following command to launch the visualizer:

docker run -it --rm  --gpus all \
-p 8501:8501 --entrypoint streamlit \ 
kartvel/deep-illuminator run streamlit/streamlit_app.py

You will be able to interact with it on localhost:8501. Note: If you do not have NVIDIA gpu support enabled for docker simply remove the --gpus all option.

Generating Variants

It is possible to quickly generate multiple variants for images contained in a directory by using the following command:

docker run -it --rm --gpus all \                                                                                               ─╯
-v /path/to/input/images:/app/probe_relighting/originals \
-v /path/to/save/directory:/app/probe_relighting/output \
kartvel/deep-illuminator --[options]

Options

Option Values Description
mode ['synthetic', 'mid'] Selecting the style of probes used as a relighting guide.
step int Increment for the granularity of relighted images. max mid: 24, max synthetic: 360

Buidling Docker image or running without a container

Please read the following for other options: instructions

Benchmarks

Improved performance of R2D2 for [email protected] on HPatches

Training Dataset Overall Viewpoint Illumination
COCO - Original 71.0 65.4 77.1
COCO - Augmented 72.2 (+1.7%) 65.7 (+0.4%) 79.2 (+2.7%)
VIDIT - Original 66.7 60.5 73.4
VIDIT - Augmented 69.2 (+3.8%) 60.9 (+0.6%) 78.1 (+6.4%)
Aachen - Original 69.4 64.1 75.0
Aachen - Augmented 72.6 (+4.6%) 66.1 (+3.1%) 79.6 (+6.1%)

Improved performance of R2D2 for the Long-Term Visual Localization challenge on Aachen v1.1

Training Dataset 0.25m, 2° 0.5m, 5° 5m, 10°
COCO - Original 62.3 77.0 79.5
COCO - Augmented 65.4 (+5.0%) 83.8 (+8.8%) 92.7 (+16%)
VIDIT - Original 40.8 53.4 61.3
VIDIT - Augmented 53.9 (+32%) 71.2 (+33%) 83.2(+36%)
Aachen - Original 60.7 72.8 83.8
Aachen - Augmented 63.4 (+4.4%) 81.7 (+12%) 92.1 (+9.9%)

Acknowledgment

The developpement of the VAE for the visualizer was made possible by the PyTorch-VAE repository.

Bibtex

If you use this code in your project, please consider citing the following paper:

@misc{chogovadze2021controllable,
      title={Controllable Data Augmentation Through Deep Relighting}, 
      author={George Chogovadze and Rémi Pautrat and Marc Pollefeys},
      year={2021},
      eprint={2110.13996},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}
sense-py-AnishaBaishya created by GitHub Classroom

Compute Statistics Here we compute statistics for a bunch of numbers. This project uses the unittest framework to test functionality. Pass the tests T

1 Oct 21, 2021
Monk is a low code Deep Learning tool and a unified wrapper for Computer Vision.

Monk - A computer vision toolkit for everyone Why use Monk Issue: Want to begin learning computer vision Solution: Start with Monk's hands-on study ro

Tessellate Imaging 507 Dec 04, 2022
A public available dataset for road boundary detection in aerial images

Topo-boundary This is the official github repo of paper Topo-boundary: A Benchmark Dataset on Topological Road-boundary Detection Using Aerial Images

Zhenhua Xu 79 Jan 04, 2023
A curated list of neural rendering resources.

Awesome-of-Neural-Rendering A curated list of neural rendering and related resources. Please feel free to pull requests or open an issue to add papers

Zhiwei ZHANG 43 Dec 09, 2022
PyTorch implementation of saliency map-aided GAN for Auto-demosaic+denosing

Saiency Map-aided GAN for RAW2RGB Mapping The PyTorch implementations and guideline for Saiency Map-aided GAN for RAW2RGB Mapping. 1 Implementations B

Yuzhi ZHAO 20 Oct 24, 2022
Reference implementation for Deep Unsupervised Learning using Nonequilibrium Thermodynamics

Diffusion Probabilistic Models This repository provides a reference implementation of the method described in the paper: Deep Unsupervised Learning us

Jascha Sohl-Dickstein 238 Jan 02, 2023
This is a tensorflow-based rotation detection benchmark, also called AlphaRotate.

AlphaRotate: A Rotation Detection Benchmark using TensorFlow Abstract AlphaRotate is maintained by Xue Yang with Shanghai Jiao Tong University supervi

yangxue 972 Jan 05, 2023
[CVPR 2021] Rethinking Text Segmentation: A Novel Dataset and A Text-Specific Refinement Approach

Rethinking Text Segmentation: A Novel Dataset and A Text-Specific Refinement Approach This is the repo to host the dataset TextSeg and code for TexRNe

SHI Lab 174 Dec 19, 2022
Expert Finding in Legal Community Question Answering

Expert Finding in Legal Community Question Answering Arian Askari, Suzan Verberne, and Gabriella Pasi. Expert Finding in Legal Community Question Answ

Arian Askari 3 Oct 31, 2022
Synthetic structured data generators

Join us on What is Synthetic Data? Synthetic data is artificially generated data that is not collected from real world events. It replicates the stati

YData 850 Jan 07, 2023
A unified 3D Transformer Pipeline for visual synthesis

Overview This is the official repo for the paper: "NÜWA: Visual Synthesis Pre-training for Neural visUal World creAtion". NÜWA is a unified multimodal

Microsoft 2.6k Jan 03, 2023
Code corresponding to The Introspective Agent: Interdependence of Strategy, Physiology, and Sensing for Embodied Agents

The Introspective Agent: Interdependence of Strategy, Physiology, and Sensing for Embodied Agents This is the code corresponding to The Introspective

0 Jan 10, 2022
Orthogonal Over-Parameterized Training

The inductive bias of a neural network is largely determined by the architecture and the training algorithm. To achieve good generalization, how to effectively train a neural network is of great impo

Weiyang Liu 11 Apr 18, 2022
给yolov5加个gui界面,使用pyqt5,yolov5是5.0版本

博文地址 https://xugaoxiang.com/2021/06/30/yolov5-pyqt5 代码执行 项目中使用YOLOv5的v5.0版本,界面文件是project.ui pip install -r requirements.txt python main.py 图片检测 视频检测

Xu GaoXiang 215 Dec 30, 2022
The Pytorch code of "Joint Distribution Matters: Deep Brownian Distance Covariance for Few-Shot Classification", CVPR 2022 (Oral).

DeepBDC for few-shot learning        Introduction In this repo, we provide the implementation of the following paper: "Joint Distribution Matters: Dee

FeiLong 116 Dec 19, 2022
Deep Learning for Natural Language Processing SS 2021 (TU Darmstadt)

Deep Learning for Natural Language Processing SS 2021 (TU Darmstadt) Task Training huge unsupervised deep neural networks yields to strong progress in

2 Aug 05, 2022
Improving Convolutional Networks via Attention Transfer (ICLR 2017)

Attention Transfer PyTorch code for "Paying More Attention to Attention: Improving the Performance of Convolutional Neural Networks via Attention Tran

Sergey Zagoruyko 1.4k Dec 23, 2022
Solutions of Reinforcement Learning 2nd Edition

Solutions of Reinforcement Learning, An Introduction

YIFAN WANG 1.4k Dec 30, 2022
AFL binary instrumentation

E9AFL --- Binary AFL E9AFL inserts American Fuzzy Lop (AFL) instrumentation into x86_64 Linux binaries. This allows binaries to be fuzzed without the

242 Dec 12, 2022
ICON: Implicit Clothed humans Obtained from Normals

ICON: Implicit Clothed humans Obtained from Normals arXiv, December 2021. Yuliang Xiu · Jinlong Yang · Dimitrios Tzionas · Michael J. Black Table of C

Yuliang Xiu 1.1k Dec 30, 2022