qubvel / albumentations

fast image augmentation library and easy to use wrapper around other libraries

Albumentations

• The library is faster than other libraries on most of the transformations.
• Based on numpy, OpenCV, imgaug picking the best from each of them.
• Simple, flexible API that allows the library to be used in any computer vision pipeline.
• Large, diverse set of transformations.
• Easy to extend the library to wrap around other libraries.
• Easy to extend to other tasks.
• Supports transformations on images, masks, key points and bounding boxes.
• Supports python 2.7-3.7
• Easy integration with PyTorch.
• Easy transfer from torchvision.
• Was used to get top results in many DL competitions at Kaggle, topcoder, CVPR, MICCAI.
• Written by Kaggle Masters.

Table of contents

• How to use
• Authors
• Installation
  • PyPI
  • Conda
• Documentation
• Pixel-level transforms
• Spatial-level transforms
• Migrating from torchvision to albumentations
• Benchmarking results
• Contributing
  • Adding new transforms
• Building the documentation
• Comments
• Citing

How to use

All in one showcase notebook - showcase.ipynb

Classification - example.ipynb

Object detection - example_bboxes.ipynb

Non-8-bit images - example_16_bit_tiff.ipynb

Image segmentation example_kaggle_salt.ipynb

Keypoints example_keypoints.ipynb

Custom targets example_multi_target.ipynb

Weather transforms example_weather_transforms.ipynb

Serialization serialization.ipynb

You can use this Google Colaboratory notebook to adjust image augmentation parameters and see the resulting images.

Authors

Alexander Buslaev

Alex Parinov

Vladimir I. Iglovikov

Evegene Khvedchenya

Installation

PyPI

You can use pip to install albumentations:

pip install albumentations

If you want to get the latest version of the code before it is released on PyPI you can install the library from GitHub:

pip install -U git+https://github.com/albu/albumentations

And it also works in Kaggle GPU kernels (proof)

!pip install albumentations > /dev/null

Conda

To install albumentations using conda we need first to install imgaug via conda-forge collection

conda install -c conda-forge imgaug
conda install albumentations -c albumentations

Documentation

The full documentation is available at albumentations.readthedocs.io.

Pixel-level transforms

Pixel-level transforms will change just an input image and will leave any additional targets such as masks, bounding boxes, and keypoints unchanged. The list of pixel-level transforms:

• Blur
• CLAHE
• ChannelDropout
• ChannelShuffle
• CoarseDropout
• Cutout
• Equalize
• FromFloat
• GaussNoise
• GaussianBlur
• HueSaturationValue
• IAAAdditiveGaussianNoise
• IAAEmboss
• IAASharpen
• IAASuperpixels
• ISONoise
• InvertImg
• JpegCompression
• MedianBlur
• MotionBlur
• Normalize
• RGBShift
• RandomBrightness
• RandomBrightnessContrast
• RandomContrast
• RandomFog
• RandomGamma
• RandomRain
• RandomShadow
• RandomSnow
• RandomSunFlare
• Solarize
• ToFloat
• ToGray

Spatial-level transforms

Spatial-level transforms will simultaneously change both an input image as well as additional targets such as masks, bounding boxes, and keypoints. The following table shows which additional targets are supported by each transform.

Transform	Image	Masks	BBoxes	Keypoints
CenterCrop	✓	✓	✓	✓
Crop	✓	✓	✓
ElasticTransform	✓	✓
Flip	✓	✓	✓	✓
GridDistortion	✓	✓
HorizontalFlip	✓	✓	✓	✓
IAAAffine	✓	✓	✓	✓
IAACropAndPad	✓	✓	✓	✓
IAAFliplr	✓	✓	✓	✓
IAAFlipud	✓	✓	✓	✓
IAAPerspective	✓	✓	✓	✓
IAAPiecewiseAffine	✓	✓	✓	✓
Lambda	✓	✓	✓	✓
LongestMaxSize	✓	✓	✓
NoOp	✓	✓	✓	✓
OpticalDistortion	✓	✓
PadIfNeeded	✓	✓	✓	✓
RandomCrop	✓	✓	✓	✓
RandomCropNearBBox	✓	✓	✓
RandomGridShuffle	✓	✓
RandomResizedCrop	✓	✓	✓	✓
RandomRotate90	✓	✓	✓	✓
RandomScale	✓	✓	✓	✓
RandomSizedBBoxSafeCrop	✓	✓	✓
RandomSizedCrop	✓	✓	✓	✓
Resize	✓	✓	✓
Rotate	✓	✓	✓	✓
ShiftScaleRotate	✓	✓	✓	✓
SmallestMaxSize	✓	✓	✓
Transpose	✓	✓	✓
VerticalFlip	✓	✓	✓	✓

Migrating from torchvision to albumentations

Migrating from torchvision to albumentations is simple - you just need to change a few lines of code. Albumentations has equivalents for common torchvision transforms as well as plenty of transforms that are not presented in torchvision. migrating_from_torchvision_to_albumentations.ipynb shows how one can migrate code from torchvision to albumentations.

Benchmarking results

To run the benchmark yourself follow the instructions in benchmark/README.md

Results for running the benchmark on first 2000 images from the ImageNet validation set using an Intel Xeon Gold 6140 CPU. The table shows how many images per second can be processed on a single core, higher is better.

	albumentations 0.3.0	imgaug 0.2.9	torchvision (Pillow backend) 0.3.0	torchvision (Pillow-SIMD backend) 0.3.0	Keras 2.2.4	Augmentor 0.2.3	solt 0.1.6
RandomCrop64	271641	3373	26538	83251	-	22535	21383
PadToSize512	2818	-	414	422	-	-	2539
Resize512	2168	696	296	1046	-	297	1907
HorizontalFlip	991	162	4881	4595	167	4595	166
VerticalFlip	4244	2278	3066	3598	4162	2985	3486
Rotate	702	475	82	110	7	37	239
ShiftScaleRotate	1761	761	81	105	9	-	-
Brightness	652	1520	316	408	149	308	2506
Contrast	1059	1539	231	316	-	241	2524
BrightnessContrast	646	861	128	174	-	130	1305
ShiftHSV	168	208	32	42	-	-	107
ShiftRGB	555	1521	-	-	477	-	-
Gamma	483	-	801	858	-	-	447
Grayscale	12097	88	728	915	-	2182	9140

Python and library versions: Python 3.7.3 | Anaconda, numpy 1.16.4, pillow 6.0.0, pillow-simd 5.3.0.post1, opencv-python 4.1.0.25, scikit-image 0.15.0, scipy 1.3.0.

Contributing

1. Clone the repository:

   git clone [e-mail]:albu/albumentations.git
   cd albumentations
   

2. Install the library in development mode:

   pip install -e .[tests]
   

3. Run tests:

   pytest
   

4. Run flake8 to perform PEP8 and PEP257 style checks and to check code for lint errors.

   flake8
   

Adding new transforms

If you are contributing a new transformation, make sure to update "Pixel-level transforms" or/and "Spatial-level transforms" sections of this file (README.md). To do this, simply run (with python3 only):

python3 tools/make_transforms_docs.py make

and copy/paste the results in the corresponding sections. To validate your modifications, you can run:

python3 tools/make_transforms_docs.py check README.md

Building the documentation

1. Go to docs/ directory

   cd docs
   

2. Install required libraries

   pip install -r requirements.txt
   

3. Build html files

   make html
   

4. Open _build/html/index.html in browser.

Alternatively, you can start a web server that rebuilds the documentation automatically when a change is detected by running make livehtml

Comments

In some systems, in the multiple GPU regime PyTorch may deadlock the DataLoader if OpenCV was compiled with OpenCL optimizations. Adding the following two lines before the library import may help. For more details pytorch/pytorch#1355

cv2.setNumThreads(0)
cv2.ocl.setUseOpenCL(False)

Citing

If you find this library useful for your research, please consider citing:

@article{2018arXiv180906839B,
    author = {A. Buslaev, A. Parinov, E. Khvedchenya, V.~I. Iglovikov and A.~A. Kalinin},
     title = "{Albumentations: fast and flexible image augmentations}",
   journal = {ArXiv e-prints},
    eprint = {1809.06839},
      year = 2018      
}