This library helps you with augmenting images for your machine learning projects. It converts a set of input images into a new, much larger set of slightly altered images. Many very popular projects have been integrated. New methods like augmix,cutmix,are being tracked. Whether you're a researcher or an engineer, just enjoy it!
Popular Projects
imgaug
- intro: 2019
- github star: 7.8k
- github: https://github.com/aleju/imgaug
Albumentations
Albumentations: fast and flexible image augmentations
- intro: ArXiv 2018
- github star: 4.1k
- arxiv: https://arxiv.org/abs/1809.06839v1
- github: https://github.com/albumentations-team/albumentations
Augmentor
Biomedical image augmentation using Augmentor
- intro: Bioinformatics
- github star: 3.7k
- arxiv: https://github.com/mdbloice/Augmentor
- github: https://github.com/mdbloice/Augmentor
- docs: https://augmentor.readthedocs.io/en/master/
Augmentor is a Python package designed to aid the augmentation and artificial generation of image data for machine learning tasks. It is primarily a data augmentation tool, but will also incorporate basic image pre-processing functionality.
Papers&Codes
mixup
Mixup: BEYOND EMPIRICAL RISK MINIMIZATION
- intro: ICLR2018
- arxiv: https://arxiv.org/abs/1710.09412
- github: https://github.com/facebookresearch/mixup-cifar10
Mixup is a generic and straightforward data augmentation principle. In essence, mixup trains a neural network on convex combinations of pairs of examples and their labels. By doing so, mixup regularizes the neural network to favor simple linear behavior in-between training examples.
Cutout
Improved Regularization of Convolutional Neural Networks with Cutout
- intro: arXiv 2017
- arxiv: https://arxiv.org/abs/1708.04552
- github: https://github.com/uoguelph-mlrg/Cutout
Cutmix
CutMix:Regularization Strategy to Train Strong Classifiers with Localizable Features
- intro: ICCV 2019 (oral talk)
- arxiv: https://arxiv.org/pdf/1905.04899.pdf
- github: https://github.com/clovaai/CutMix-PyTorch
Augmix
AUGMIX: A SIMPLE DATA PROCESSING METHOD TO IMPROVE ROBUSTNESS AND UNCERTAINTY
- intro: ICLR 2020
- arxiv: https://arxiv.org/pdf/1912.02781.pdf
- github: https://github.com/google-research/augmix
fast-autoaugment
Fast AutoAugment
- intro: NeurIPS 2019
- github star: 671
- arxiv: https://arxiv.org/abs/1905.00397
- github: https://github.com/kakaobrain/fast-autoaugment
AutoAugment
AutoAugment:Learning Augmentation Strategies from Data
- intro: CVPR 2019
- provider: google
- arxiv: https://arxiv.org/pdf/1805.09501v3.pdf
- github: https://github.com/DeepVoltaire/AutoAugment
RandAugment
RandAugment: Practical automated data augmentation with a reduced search space
- intro: ICLR 2020
- provider: google
- arxiv: https://arxiv.org/pdf/1912.02781.pdf
- github: https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet
GridMask
GridMaskDataAugmentation
- intro: 2020.01
- arxiv: https://arxiv.org/abs/2001.04086
- github: https://github.com/akuxcw/GridMask
imagecorruptions
Benchmarking Robustness in Object Detection:Autonomous Driving when Winter is Coming
- intro: arXiv 2019
- arxiv: <://arxiv.org/pdf/1912.02781.pdf>
- provider: UC Berkeley
- github: https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix
- github: https://github.com/junyanz/CycleGAN
CycleGAN
Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networkss
- intro: ICCV 2017
- arxiv: https://arxiv.org/pdf/1912.02781.pdf
- provider: UC Berkeley
- github: https://github.com/CrazyVertigo/imagecorruptions
Small Object Augmentation
Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networkss
- intro: 2017
- arxiv: https://arxiv.org/pdf/1902.07296.pdf
- github: https://github.com/gmayday1997/SmallObjectAugmentation
Continuous updating...
If you find this library useful for your research, please consider starring the GitHub repository.