Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/albumentations-team/albumentations
Fast image augmentation library and an easy-to-use wrapper around other libraries. Documentation: https://albumentations.ai/docs/ Paper about the library: https://www.mdpi.com/2078-2489/11/2/125
https://github.com/albumentations-team/albumentations
augmentation deep-learning detection fast-augmentations image-augmentation image-classification image-processing image-segmentation machine-learning object-detection python segmentation
Last synced: 11 days ago
JSON representation
Fast image augmentation library and an easy-to-use wrapper around other libraries. Documentation: https://albumentations.ai/docs/ Paper about the library: https://www.mdpi.com/2078-2489/11/2/125
- Host: GitHub
- URL: https://github.com/albumentations-team/albumentations
- Owner: albumentations-team
- License: mit
- Created: 2018-06-06T03:10:50.000Z (over 6 years ago)
- Default Branch: main
- Last Pushed: 2024-05-08T23:00:15.000Z (6 months ago)
- Last Synced: 2024-05-11T10:00:42.138Z (6 months ago)
- Topics: augmentation, deep-learning, detection, fast-augmentations, image-augmentation, image-classification, image-processing, image-segmentation, machine-learning, object-detection, python, segmentation
- Language: Python
- Homepage: https://albumentations.ai
- Size: 97.9 MB
- Stars: 13,470
- Watchers: 128
- Forks: 1,594
- Open Issues: 424
-
Metadata Files:
- Readme: README.md
- Contributing: CONTRIBUTING.md
- Funding: .github/FUNDING.yml
- License: LICENSE
Awesome Lists containing this project
- awesome-python-machine-learning-resources - GitHub - 41% open · ⏱️ 24.08.2022): (图像数据与CV)
- Algorithms-Cheatsheet-Resources - Fast image augmentation library and easy to use wrapper around other libraries
- awesome-open-data-centric-ai - Albumentations - to-use wrapper around other libraries. |  | <a href="https://github.com/albumentations-team/albumentations/blob/main/LICENSE"><img src="https://img.shields.io/github/license/albumentations-team/albumentations" height="15"/></a> | (Augmentation and synthetic data)
- awesome-cv - albumentations - to-use wrapper around other libraries | (Repos / Tags: Object Classification `[ObjCls]`, Object Detection `[ObjDet]`, Object Segmentation `[ObjSeg]`, General Library `[GenLib]`, Text Reading / Object Character Recognition `[OCR]`, Action Recognition `[ActRec]`, Object Tracking `[ObjTrk]`, Data Augmentation `[DatAug]`, Simultaneous Localization and Mapping `[SLAM]`, Outlier/Anomaly/Novelty Detection `[NvlDet]`, Content-based Image Retrieval `[CBIR]`, Image Enhancement `[ImgEnh]`, Aesthetic Assessment `[AesAss]`, Explainable Artificial Intelligence `[XAI]`, Text-to-Image Generation `[TexImg]`, Pose Estimation `[PosEst]`, Video Matting `[VidMat]`, Eye Tracking `[EyeTrk]`)
- awesome-object-detection-datasets - Albumentations - team/albumentations?style=social"/> : Albumentations is a Python library for image augmentation. Image augmentation is used in deep learning and computer vision tasks to increase the quality of trained models. The purpose of image augmentation is to create new training samples from the existing data. "Albumentations: Fast and Flexible Image Augmentations". (**[Information 2020](https://www.mdpi.com/2078-2489/11/2/125)**) (Summary)
- awesome-list - Albumentations - A Python library for image augmentation. (Data Processing / Data Pre-processing & Loading)
- StarryDivineSky - albumentations-team/albumentations
- awesome-yolo-object-detection - Albumentations - team/albumentations?style=social"/> : Albumentations is a Python library for image augmentation. Image augmentation is used in deep learning and computer vision tasks to increase the quality of trained models. The purpose of image augmentation is to create new training samples from the existing data. "Albumentations: Fast and Flexible Image Augmentations". (**[Information 2020](https://www.mdpi.com/2078-2489/11/2/125)**) (Extensional Frameworks)
- awesome-yolo-object-detection - Albumentations - team/albumentations?style=social"/> : Albumentations is a Python library for image augmentation. Image augmentation is used in deep learning and computer vision tasks to increase the quality of trained models. The purpose of image augmentation is to create new training samples from the existing data. "Albumentations: Fast and Flexible Image Augmentations". (**[Information 2020](https://www.mdpi.com/2078-2489/11/2/125)**) (Extensional Frameworks)
README
# Albumentations
[](https://badge.fury.io/py/albumentations)
[](
https://pypi.org/project/albumentations/)
[](
https://anaconda.org/conda-forge/albumentations)
[](
https://stackoverflow.com/questions/tagged/albumentations)
[](https://opensource.org/licenses/MIT)
[Docs](https://albumentations.ai/docs/) | [Discord](https://discord.gg/AKPrrDYNAt) | [Twitter](https://twitter.com/albumentations) | [LinkedIn](https://www.linkedin.com/company/100504475/)
Albumentations is a Python library for image augmentation. Image augmentation is used in deep learning and computer vision tasks to increase the quality of trained models. The purpose of image augmentation is to create new training samples from the existing data.
Here is an example of how you can apply some [pixel-level](#pixel-level-transforms) augmentations from Albumentations to create new images from the original one:
## Why Albumentations
- Albumentations **[supports all common computer vision tasks](#i-want-to-use-albumentations-for-the-specific-task-such-as-classification-or-segmentation)** such as classification, semantic segmentation, instance segmentation, object detection, and pose estimation.
- The library provides **[a simple unified API](#a-simple-example)** to work with all data types: images (RBG-images, grayscale images, multispectral images), segmentation masks, bounding boxes, and keypoints.
- The library contains **[more than 70 different augmentations](#list-of-augmentations)** to generate new training samples from the existing data.
- Albumentations is [**fast**](#benchmarking-results). We benchmark each new release to ensure that augmentations provide maximum speed.
- It **[works with popular deep learning frameworks](#i-want-to-know-how-to-use-albumentations-with-deep-learning-frameworks)** such as PyTorch and TensorFlow. By the way, Albumentations is a part of the [PyTorch ecosystem](https://pytorch.org/ecosystem/).
- [**Written by experts**](#authors). The authors have experience both working on production computer vision systems and participating in competitive machine learning. Many core team members are Kaggle Masters and Grandmasters.
- The library is [**widely used**](#who-is-using-albumentations) in industry, deep learning research, machine learning competitions, and open source projects.
## Sponsors
## Table of contents
- [Albumentations](#albumentations)
- [Why Albumentations](#why-albumentations)
- [Sponsors](#sponsors)
- [Table of contents](#table-of-contents)
- [Authors](#authors)
- [Current Maintainer](#current-maintainer)
- [Emeritus Core Team Members](#emeritus-core-team-members)
- [Installation](#installation)
- [Documentation](#documentation)
- [A simple example](#a-simple-example)
- [Getting started](#getting-started)
- [I am new to image augmentation](#i-am-new-to-image-augmentation)
- [I want to use Albumentations for the specific task such as classification or segmentation](#i-want-to-use-albumentations-for-the-specific-task-such-as-classification-or-segmentation)
- [I want to know how to use Albumentations with deep learning frameworks](#i-want-to-know-how-to-use-albumentations-with-deep-learning-frameworks)
- [I want to explore augmentations and see Albumentations in action](#i-want-to-explore-augmentations-and-see-albumentations-in-action)
- [Who is using Albumentations](#who-is-using-albumentations)
- [See also](#see-also)
- [List of augmentations](#list-of-augmentations)
- [Pixel-level transforms](#pixel-level-transforms)
- [Spatial-level transforms](#spatial-level-transforms)
- [A few more examples of **augmentations**](#a-few-more-examples-of-augmentations)
- [Semantic segmentation on the Inria dataset](#semantic-segmentation-on-the-inria-dataset)
- [Medical imaging](#medical-imaging)
- [Object detection and semantic segmentation on the Mapillary Vistas dataset](#object-detection-and-semantic-segmentation-on-the-mapillary-vistas-dataset)
- [Keypoints augmentation](#keypoints-augmentation)
- [Benchmarking results](#benchmarking-results)
- [Contributing](#contributing)
- [Community and Support](#community-and-support)
- [Comments](#comments)
- [Citing](#citing)
## Authors
### Current Maintainer
[**Vladimir I. Iglovikov**](https://www.linkedin.com/in/iglovikov/) | [Kaggle Grandmaster](https://www.kaggle.com/iglovikov)
### Emeritus Core Team Members
[**Mikhail Druzhinin**](https://www.linkedin.com/in/mikhail-druzhinin-548229100/) | [Kaggle Expert](https://www.kaggle.com/dipetm)
[**Alex Parinov**](https://www.linkedin.com/in/alex-parinov/) | [Kaggle Master](https://www.kaggle.com/creafz)
[**Alexander Buslaev**](https://www.linkedin.com/in/al-buslaev/) | [Kaggle Master](https://www.kaggle.com/albuslaev)
[**Eugene Khvedchenya**](https://www.linkedin.com/in/cvtalks/) | [Kaggle Grandmaster](https://www.kaggle.com/bloodaxe)
## Installation
Albumentations requires Python 3.9 or higher. To install the latest version from PyPI:
```bash
pip install -U albumentations
```
Other installation options are described in the [documentation](https://albumentations.ai/docs/getting_started/installation/).
## Documentation
The full documentation is available at **[https://albumentations.ai/docs/](https://albumentations.ai/docs/)**.
## A simple example
```python
import albumentations as A
import cv2
# Declare an augmentation pipeline
transform = A.Compose([
A.RandomCrop(width=256, height=256),
A.HorizontalFlip(p=0.5),
A.RandomBrightnessContrast(p=0.2),
])
# Read an image with OpenCV and convert it to the RGB colorspace
image = cv2.imread("image.jpg")
image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
# Augment an image
transformed = transform(image=image)
transformed_image = transformed["image"]
```
## Getting started
### I am new to image augmentation
Please start with the [introduction articles](https://albumentations.ai/docs/#introduction-to-image-augmentation) about why image augmentation is important and how it helps to build better models.
### I want to use Albumentations for the specific task such as classification or segmentation
If you want to use Albumentations for a specific task such as classification, segmentation, or object detection, refer to the [set of articles](https://albumentations.ai/docs/#getting-started-with-albumentations) that has an in-depth description of this task. We also have a [list of examples](https://albumentations.ai/docs/examples/) on applying Albumentations for different use cases.
### I want to know how to use Albumentations with deep learning frameworks
We have [examples of using Albumentations](https://albumentations.ai/docs/#examples-of-how-to-use-albumentations-with-different-deep-learning-frameworks) along with PyTorch and TensorFlow.
### I want to explore augmentations and see Albumentations in action
Check the [online demo of the library](https://albumentations-demo.herokuapp.com/). With it, you can apply augmentations to different images and see the result. Also, we have a [list of all available augmentations and their targets](#list-of-augmentations).
## Who is using Albumentations
### See also
- [A list of papers that cite Albumentations](https://scholar.google.com/citations?view_op=view_citation&citation_for_view=vkjh9X0AAAAJ:r0BpntZqJG4C).
- [A list of teams that were using Albumentations and took high places in machine learning competitions](https://albumentations.ai/whos_using#competitions).
- [Open source projects that use Albumentations](https://github.com/albumentations-team/albumentations/network/dependents?dependent_type=PACKAGE).
## List of augmentations
### 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:
- [AdvancedBlur](https://explore.albumentations.ai/transform/AdvancedBlur)
- [Blur](https://explore.albumentations.ai/transform/Blur)
- [CLAHE](https://explore.albumentations.ai/transform/CLAHE)
- [ChannelDropout](https://explore.albumentations.ai/transform/ChannelDropout)
- [ChannelShuffle](https://explore.albumentations.ai/transform/ChannelShuffle)
- [ChromaticAberration](https://explore.albumentations.ai/transform/ChromaticAberration)
- [ColorJitter](https://explore.albumentations.ai/transform/ColorJitter)
- [Defocus](https://explore.albumentations.ai/transform/Defocus)
- [Downscale](https://explore.albumentations.ai/transform/Downscale)
- [Emboss](https://explore.albumentations.ai/transform/Emboss)
- [Equalize](https://explore.albumentations.ai/transform/Equalize)
- [FDA](https://explore.albumentations.ai/transform/FDA)
- [FancyPCA](https://explore.albumentations.ai/transform/FancyPCA)
- [FromFloat](https://explore.albumentations.ai/transform/FromFloat)
- [GaussNoise](https://explore.albumentations.ai/transform/GaussNoise)
- [GaussianBlur](https://explore.albumentations.ai/transform/GaussianBlur)
- [GlassBlur](https://explore.albumentations.ai/transform/GlassBlur)
- [HistogramMatching](https://explore.albumentations.ai/transform/HistogramMatching)
- [HueSaturationValue](https://explore.albumentations.ai/transform/HueSaturationValue)
- [ISONoise](https://explore.albumentations.ai/transform/ISONoise)
- [ImageCompression](https://explore.albumentations.ai/transform/ImageCompression)
- [InvertImg](https://explore.albumentations.ai/transform/InvertImg)
- [MedianBlur](https://explore.albumentations.ai/transform/MedianBlur)
- [MotionBlur](https://explore.albumentations.ai/transform/MotionBlur)
- [MultiplicativeNoise](https://explore.albumentations.ai/transform/MultiplicativeNoise)
- [Normalize](https://explore.albumentations.ai/transform/Normalize)
- [PixelDistributionAdaptation](https://explore.albumentations.ai/transform/PixelDistributionAdaptation)
- [PlanckianJitter](https://explore.albumentations.ai/transform/PlanckianJitter)
- [Posterize](https://explore.albumentations.ai/transform/Posterize)
- [RGBShift](https://explore.albumentations.ai/transform/RGBShift)
- [RandomBrightnessContrast](https://explore.albumentations.ai/transform/RandomBrightnessContrast)
- [RandomFog](https://explore.albumentations.ai/transform/RandomFog)
- [RandomGamma](https://explore.albumentations.ai/transform/RandomGamma)
- [RandomGravel](https://explore.albumentations.ai/transform/RandomGravel)
- [RandomRain](https://explore.albumentations.ai/transform/RandomRain)
- [RandomShadow](https://explore.albumentations.ai/transform/RandomShadow)
- [RandomSnow](https://explore.albumentations.ai/transform/RandomSnow)
- [RandomSunFlare](https://explore.albumentations.ai/transform/RandomSunFlare)
- [RandomToneCurve](https://explore.albumentations.ai/transform/RandomToneCurve)
- [RingingOvershoot](https://explore.albumentations.ai/transform/RingingOvershoot)
- [Sharpen](https://explore.albumentations.ai/transform/Sharpen)
- [Solarize](https://explore.albumentations.ai/transform/Solarize)
- [Spatter](https://explore.albumentations.ai/transform/Spatter)
- [Superpixels](https://explore.albumentations.ai/transform/Superpixels)
- [TemplateTransform](https://explore.albumentations.ai/transform/TemplateTransform)
- [TextImage](https://explore.albumentations.ai/transform/TextImage)
- [ToFloat](https://explore.albumentations.ai/transform/ToFloat)
- [ToGray](https://explore.albumentations.ai/transform/ToGray)
- [ToRGB](https://explore.albumentations.ai/transform/ToRGB)
- [ToSepia](https://explore.albumentations.ai/transform/ToSepia)
- [UnsharpMask](https://explore.albumentations.ai/transform/UnsharpMask)
- [ZoomBlur](https://explore.albumentations.ai/transform/ZoomBlur)
### 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 | Mask | BBoxes | Keypoints |
| ------------------------------------------------------------------------------------------------ | :---: | :--: | :----: | :-------: |
| [Affine](https://explore.albumentations.ai/transform/Affine) | ✓ | ✓ | ✓ | ✓ |
| [BBoxSafeRandomCrop](https://explore.albumentations.ai/transform/BBoxSafeRandomCrop) | ✓ | ✓ | ✓ | ✓ |
| [CenterCrop](https://explore.albumentations.ai/transform/CenterCrop) | ✓ | ✓ | ✓ | ✓ |
| [CoarseDropout](https://explore.albumentations.ai/transform/CoarseDropout) | ✓ | ✓ | ✓ | ✓ |
| [Crop](https://explore.albumentations.ai/transform/Crop) | ✓ | ✓ | ✓ | ✓ |
| [CropAndPad](https://explore.albumentations.ai/transform/CropAndPad) | ✓ | ✓ | ✓ | ✓ |
| [CropNonEmptyMaskIfExists](https://explore.albumentations.ai/transform/CropNonEmptyMaskIfExists) | ✓ | ✓ | ✓ | ✓ |
| [D4](https://explore.albumentations.ai/transform/D4) | ✓ | ✓ | ✓ | ✓ |
| [ElasticTransform](https://explore.albumentations.ai/transform/ElasticTransform) | ✓ | ✓ | ✓ | ✓ |
| [GridDistortion](https://explore.albumentations.ai/transform/GridDistortion) | ✓ | ✓ | ✓ | ✓ |
| [GridDropout](https://explore.albumentations.ai/transform/GridDropout) | ✓ | ✓ | ✓ | ✓ |
| [GridElasticDeform](https://explore.albumentations.ai/transform/GridElasticDeform) | ✓ | ✓ | ✓ | ✓ |
| [HorizontalFlip](https://explore.albumentations.ai/transform/HorizontalFlip) | ✓ | ✓ | ✓ | ✓ |
| [Lambda](https://explore.albumentations.ai/transform/Lambda) | ✓ | ✓ | ✓ | ✓ |
| [LongestMaxSize](https://explore.albumentations.ai/transform/LongestMaxSize) | ✓ | ✓ | ✓ | ✓ |
| [MaskDropout](https://explore.albumentations.ai/transform/MaskDropout) | ✓ | ✓ | ✓ | ✓ |
| [Morphological](https://explore.albumentations.ai/transform/Morphological) | ✓ | ✓ | ✓ | ✓ |
| [NoOp](https://explore.albumentations.ai/transform/NoOp) | ✓ | ✓ | ✓ | ✓ |
| [OpticalDistortion](https://explore.albumentations.ai/transform/OpticalDistortion) | ✓ | ✓ | ✓ | ✓ |
| [OverlayElements](https://explore.albumentations.ai/transform/OverlayElements) | ✓ | ✓ | | |
| [PadIfNeeded](https://explore.albumentations.ai/transform/PadIfNeeded) | ✓ | ✓ | ✓ | ✓ |
| [Perspective](https://explore.albumentations.ai/transform/Perspective) | ✓ | ✓ | ✓ | ✓ |
| [PiecewiseAffine](https://explore.albumentations.ai/transform/PiecewiseAffine) | ✓ | ✓ | ✓ | ✓ |
| [PixelDropout](https://explore.albumentations.ai/transform/PixelDropout) | ✓ | ✓ | ✓ | ✓ |
| [RandomCrop](https://explore.albumentations.ai/transform/RandomCrop) | ✓ | ✓ | ✓ | ✓ |
| [RandomCropFromBorders](https://explore.albumentations.ai/transform/RandomCropFromBorders) | ✓ | ✓ | ✓ | ✓ |
| [RandomGridShuffle](https://explore.albumentations.ai/transform/RandomGridShuffle) | ✓ | ✓ | | ✓ |
| [RandomResizedCrop](https://explore.albumentations.ai/transform/RandomResizedCrop) | ✓ | ✓ | ✓ | ✓ |
| [RandomRotate90](https://explore.albumentations.ai/transform/RandomRotate90) | ✓ | ✓ | ✓ | ✓ |
| [RandomScale](https://explore.albumentations.ai/transform/RandomScale) | ✓ | ✓ | ✓ | ✓ |
| [RandomSizedBBoxSafeCrop](https://explore.albumentations.ai/transform/RandomSizedBBoxSafeCrop) | ✓ | ✓ | ✓ | ✓ |
| [RandomSizedCrop](https://explore.albumentations.ai/transform/RandomSizedCrop) | ✓ | ✓ | ✓ | ✓ |
| [Resize](https://explore.albumentations.ai/transform/Resize) | ✓ | ✓ | ✓ | ✓ |
| [Rotate](https://explore.albumentations.ai/transform/Rotate) | ✓ | ✓ | ✓ | ✓ |
| [SafeRotate](https://explore.albumentations.ai/transform/SafeRotate) | ✓ | ✓ | ✓ | ✓ |
| [ShiftScaleRotate](https://explore.albumentations.ai/transform/ShiftScaleRotate) | ✓ | ✓ | ✓ | ✓ |
| [SmallestMaxSize](https://explore.albumentations.ai/transform/SmallestMaxSize) | ✓ | ✓ | ✓ | ✓ |
| [Transpose](https://explore.albumentations.ai/transform/Transpose) | ✓ | ✓ | ✓ | ✓ |
| [VerticalFlip](https://explore.albumentations.ai/transform/VerticalFlip) | ✓ | ✓ | ✓ | ✓ |
| [XYMasking](https://explore.albumentations.ai/transform/XYMasking) | ✓ | ✓ | ✓ | ✓ |
## A few more examples of **augmentations**
### Semantic segmentation on the Inria dataset
### Medical imaging
### Object detection and semantic segmentation on the Mapillary Vistas dataset
### Keypoints augmentation
## Benchmarking results
To run the benchmark yourself, follow the instructions in [benchmark/README.md](https://github.com/albumentations-team/albumentations/blob/master/benchmark/README.md)
Results for running the benchmark on the first 2000 images from the ImageNet validation set using an AMD Ryzen Threadripper 3970X CPU.
The table shows how many images per second can be processed on a single core; higher is better.
| Library | Version |
|---------|---------|
| Python | 3.10.13 (main, Sep 11 2023, 13:44:35) [GCC 11.2.0] |
| albumentations | 1.4.11 |
| imgaug | 0.4.0 |
| torchvision | 0.18.1+rocm6.0 |
| numpy | 1.26.4 |
| opencv-python-headless | 4.10.0.84 |
| scikit-image | 0.24.0 |
| scipy | 1.14.0 |
| pillow | 10.4.0 |
| kornia | 0.7.3 |
| augly | 1.0.0 |
| |albumentations
1.4.11|torchvision
0.18.1+rocm6.0|kornia
0.7.3|augly
1.0.0|imgaug
0.4.0|
|-----------------|--------------------------------------|--------------------------------------------|------------------------------|-----------------------------|------------------------------|
|HorizontalFlip |**8017 ± 12** |2436 ± 2 |935 ± 3 |3575 ± 4 |4806 ± 7 |
|VerticalFlip |7366 ± 7 |2563 ± 8 |943 ± 1 |4949 ± 5 |**8159 ± 21** |
|Rotate |570 ± 12 |152 ± 2 |207 ± 1 |**633 ± 2** |496 ± 2 |
|Affine |**1382 ± 31** |162 ± 1 |201 ± 1 |- |682 ± 2 |
|Equalize |1027 ± 2 |336 ± 2 |77 ± 1 |- |**1183 ± 1** |
|RandomCrop64 |**19986 ± 57** |15336 ± 16 |811 ± 1 |19882 ± 356 |5410 ± 5 |
|RandomResizedCrop|**2308 ± 7** |1046 ± 3 |187 ± 1 |- |- |
|ShiftRGB |1240 ± 3 |- |425 ± 2 |- |**1554 ± 6** |
|Resize |**2314 ± 9** |1272 ± 3 |201 ± 3 |431 ± 1 |1715 ± 2 |
|RandomGamma |**2552 ± 2** |232 ± 1 |211 ± 1 |- |1794 ± 1 |
|Grayscale |**7313 ± 4** |1652 ± 2 |443 ± 2 |2639 ± 2 |1171 ± 23 |
|ColorJitter |**396 ± 1** |51 ± 1 |50 ± 1 |224 ± 1 |- |
|PlankianJitter |449 ± 1 |- |**598 ± 1** |- |- |
|RandomPerspective|471 ± 1 |123 ± 1 |114 ± 1 |- |**478 ± 2** |
|GaussianBlur |**2099 ± 2** |113 ± 2 |79 ± 2 |165 ± 1 |1244 ± 2 |
|MedianBlur |538 ± 1 |- |3 ± 1 |- |**565 ± 1** |
|MotionBlur |**2197 ± 9** |- |102 ± 1 |- |508 ± 1 |
|Posterize |2449 ± 1 |**2587 ± 3** |339 ± 6 |- |1547 ± 1 |
|JpegCompression |**827 ± 1** |- |50 ± 2 |684 ± 1 |428 ± 4 |
|GaussianNoise |78 ± 1 |- |- |67 ± 1 |**128 ± 1** |
|Elastic |127 ± 1 |3 ± 1 |1 ± 1 |- |**130 ± 1** |
|Normalize |**971 ± 2** |449 ± 1 |415 ± 1 |- |- |
## Contributing
To create a pull request to the repository, follow the documentation at [CONTRIBUTING.md](CONTRIBUTING.md)
## Community and Support
- [Twitter](https://twitter.com/albumentations)
- [Discord](https://discord.gg/AKPrrDYNAt)
## 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 [https://github.com/pytorch/pytorch/issues/1355](https://github.com/pytorch/pytorch/issues/1355)
```python
cv2.setNumThreads(0)
cv2.ocl.setUseOpenCL(False)
```
## Citing
If you find this library useful for your research, please consider citing [Albumentations: Fast and Flexible Image Augmentations](https://www.mdpi.com/2078-2489/11/2/125):
```bibtex
@Article{info11020125,
AUTHOR = {Buslaev, Alexander and Iglovikov, Vladimir I. and Khvedchenya, Eugene and Parinov, Alex and Druzhinin, Mikhail and Kalinin, Alexandr A.},
TITLE = {Albumentations: Fast and Flexible Image Augmentations},
JOURNAL = {Information},
VOLUME = {11},
YEAR = {2020},
NUMBER = {2},
ARTICLE-NUMBER = {125},
URL = {https://www.mdpi.com/2078-2489/11/2/125},
ISSN = {2078-2489},
DOI = {10.3390/info11020125}
}
```