| Name | albumentations JSON |
| Version |
1.4.21
JSON |
| download |
| home_page | None |
| Summary | Fast, flexible, and advanced image augmentation library for deep learning and computer vision. Albumentations offers a wide range of transformations for images, masks, bounding boxes, and keypoints, with optimized performance and seamless integration into ML workflows. |
| upload_time | 2024-11-01 00:16:02 |
| maintainer | Vladimir Iglovikov |
| docs_url | None |
| author | Vladimir Iglovikov |
| requires_python | >=3.9 |
| license | MIT License Copyright (c) 2017 Vladimir Iglovikov, Alexander Buslaev, Alexander Parinov, Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. |
| keywords |
artificial intelligence
bounding box
computer vision
computer vision library
data augmentation
data preprocessing
deep learning
fast augmentation
image augmentation
image classification
image processing
image transformation
instance segmentation
keras
keypoint detection
machine learning
object detection
pytorch
semantic segmentation
tensorflow
|
| VCS |
|
| bugtrack_url |
|
| requirements |
No requirements were recorded.
|
| Travis-CI |
No Travis.
|
| coveralls test coverage |
No coveralls.
|
# 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 (RGB-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
<a href="https://roboflow.com/" target="_blank"><img src="https://avatars.githubusercontent.com/u/53104118?s=200&v=4" width="100"/></a>
## 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)
- [System Information](#system-information)
- [Benchmark Parameters](#benchmark-parameters)
- [Library Versions](#library-versions)
- [Performance Comparison](#performance-comparison)
- [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
<a href="https://www.apple.com/" target="_blank"><img src="https://raw.githubusercontent.com/albumentations-team/albumentations.ai/main/html/assets/img/industry/apple.jpeg" width="100"/></a>
<a href="https://research.google/" target="_blank"><img src="https://raw.githubusercontent.com/albumentations-team/albumentations.ai/main/html/assets/img/industry/google.png" width="100"/></a>
<a href="https://opensource.fb.com/" target="_blank"><img src="https://raw.githubusercontent.com/albumentations-team/albumentations.ai/main/html/assets/img/industry/meta_research.png" width="100"/></a>
<a href="https: //www.nvidia.com/en-us/research/" target="_blank"><img src="https://raw.githubusercontent.com/albumentations-team/albumentations.ai/main/html/assets/img/industry/nvidia_research.jpeg" width="100"/></a>
<a href="https://www.amazon.science/" target="_blank"><img src="https://raw.githubusercontent.com/albumentations-team/albumentations.ai/main/html/assets/img/industry/amazon_science.png" width="100"/></a>
<a href="https://opensource.microsoft.com/" target="_blank"><img src="https://raw.githubusercontent.com/albumentations-team/albumentations.ai/main/html/assets/img/industry/microsoft.png" width="100"/></a>
<a href="https://engineering.salesforce.com/open-source/" target="_blank"><img src="https://raw.githubusercontent.com/albumentations-team/albumentations.ai/main/html/assets/img/industry/salesforce_open_source.png" width="100"/></a>
<a href="https://stability.ai/" target="_blank"><img src="https://raw.githubusercontent.com/albumentations-team/albumentations.ai/main/html/assets/img/industry/stability.png" width="100"/></a>
<a href="https://www.ibm.com/opensource/" target="_blank"><img src="https://raw.githubusercontent.com/albumentations-team/albumentations.ai/main/html/assets/img/industry/ibm.jpeg" width="100"/></a>
<a href="https://huggingface.co/" target="_blank"><img src="https://raw.githubusercontent.com/albumentations-team/albumentations.ai/main/html/assets/img/industry/hugging_face.png" width="100"/></a>
<a href="https://www.sony.com/en/" target="_blank"><img src="https://raw.githubusercontent.com/albumentations-team/albumentations.ai/main/html/assets/img/industry/sony.png" width="100"/></a>
<a href="https://opensource.alibaba.com/" target="_blank"><img src="https://raw.githubusercontent.com/albumentations-team/albumentations.ai/main/html/assets/img/industry/alibaba.png" width="100"/></a>
<a href="https://opensource.tencent.com/" target="_blank"><img src="https://raw.githubusercontent.com/albumentations-team/albumentations.ai/main/html/assets/img/industry/tencent.png" width="100"/></a>
<a href="https://h2o.ai/" target="_blank"><img src="https://raw.githubusercontent.com/albumentations-team/albumentations.ai/main/html/assets/img/industry/h2o_ai.png" width="100"/></a>
### 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
<img src="https://habrastorage.org/webt/e-/6k/z-/e-6kz-fugp2heak3jzns3bc-r8o.jpeg" width=100%>
## Benchmarking results
### System Information
- Platform: macOS-15.0.1-arm64-arm-64bit
- Processor: arm
- CPU Count: 10
- Python Version: 3.12.7
### Benchmark Parameters
- Number of images: 1000
- Runs per transform: 10
- Max warmup iterations: 1000
### Library Versions
- albumentations: 1.4.20
- augly: 1.0.0
- imgaug: 0.4.0
- kornia: 0.7.3
- torchvision: 0.20.0
## Performance Comparison
| Transform | albumentations<br>1.4.20 | augly<br>1.0.0 | imgaug<br>0.4.0 | kornia<br>0.7.3 | torchvision<br>0.20.0 |
|:------------------|:---------------------------|:-----------------|:------------------|:------------------|:------------------------|
| HorizontalFlip | **8618 ± 1233** | 4807 ± 818 | 6042 ± 788 | 390 ± 106 | 914 ± 67 |
| VerticalFlip | **22847 ± 2031** | 9153 ± 1291 | 10931 ± 1844 | 1212 ± 402 | 3198 ± 200 |
| Rotate | **1146 ± 79** | 1119 ± 41 | 1136 ± 218 | 143 ± 11 | 181 ± 11 |
| Affine | 682 ± 192 | - | **774 ± 97** | 147 ± 9 | 130 ± 12 |
| Equalize | **892 ± 61** | - | 581 ± 54 | 152 ± 19 | 479 ± 12 |
| RandomCrop80 | **47341 ± 20523** | 25272 ± 1822 | 11503 ± 441 | 1510 ± 230 | 32109 ± 1241 |
| ShiftRGB | **2349 ± 76** | - | 1582 ± 65 | - | - |
| Resize | **2316 ± 166** | 611 ± 78 | 1806 ± 63 | 232 ± 24 | 195 ± 4 |
| RandomGamma | **8675 ± 274** | - | 2318 ± 269 | 108 ± 13 | - |
| Grayscale | **3056 ± 47** | 2720 ± 932 | 1681 ± 156 | 289 ± 75 | 1838 ± 130 |
| RandomPerspective | 412 ± 38 | - | **554 ± 22** | 86 ± 11 | 96 ± 5 |
| GaussianBlur | **1728 ± 89** | 242 ± 4 | 1090 ± 65 | 176 ± 18 | 79 ± 3 |
| MedianBlur | **868 ± 60** | - | 813 ± 30 | 5 ± 0 | - |
| MotionBlur | **4047 ± 67** | - | 612 ± 18 | 73 ± 2 | - |
| Posterize | **9094 ± 301** | - | 2097 ± 68 | 430 ± 49 | 3196 ± 185 |
| JpegCompression | **918 ± 23** | 778 ± 5 | 459 ± 35 | 71 ± 3 | 625 ± 17 |
| GaussianNoise | 166 ± 12 | 67 ± 2 | **206 ± 11** | 75 ± 1 | - |
| Elastic | 201 ± 5 | - | **235 ± 20** | 1 ± 0 | 2 ± 0 |
| Clahe | **454 ± 22** | - | 335 ± 43 | 94 ± 9 | - |
| CoarseDropout | **13368 ± 744** | - | 671 ± 38 | 536 ± 87 | - |
| Blur | **5267 ± 543** | 246 ± 3 | 3807 ± 325 | - | - |
| ColorJitter | **628 ± 55** | 255 ± 13 | - | 55 ± 18 | 46 ± 2 |
| Brightness | **8956 ± 300** | 1163 ± 86 | - | 472 ± 101 | 429 ± 20 |
| Contrast | **8879 ± 1426** | 736 ± 79 | - | 425 ± 52 | 335 ± 35 |
| RandomResizedCrop | **2828 ± 186** | - | - | 287 ± 58 | 511 ± 10 |
| Normalize | **1196 ± 56** | - | - | 626 ± 40 | 519 ± 12 |
| PlankianJitter | **2204 ± 385** | - | - | 813 ± 211 | - |
## 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)
- [LinkedIn](https://www.linkedin.com/company/albumentations/)
## 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}
}
```
Raw data
{
"_id": null,
"home_page": null,
"name": "albumentations",
"maintainer": "Vladimir Iglovikov",
"docs_url": null,
"requires_python": ">=3.9",
"maintainer_email": null,
"keywords": "artificial intelligence, bounding box, computer vision, computer vision library, data augmentation, data preprocessing, deep learning, fast augmentation, image augmentation, image classification, image processing, image transformation, instance segmentation, keras, keypoint detection, machine learning, object detection, pytorch, semantic segmentation, tensorflow",
"author": "Vladimir Iglovikov",
"author_email": null,
"download_url": "https://files.pythonhosted.org/packages/ac/7a/ab6f36f33cfb7639b8a1df4656624a94b03a164c0a063747cb0c59c8fdfa/albumentations-1.4.21.tar.gz",
"platform": null,
"description": "# Albumentations\n\n[](https://badge.fury.io/py/albumentations)\n\n[](\nhttps://pypi.org/project/albumentations/)\n[](\nhttps://anaconda.org/conda-forge/albumentations)\n[](\nhttps://stackoverflow.com/questions/tagged/albumentations)\n[](https://opensource.org/licenses/MIT)\n\n[Docs](https://albumentations.ai/docs/) | [Discord](https://discord.gg/AKPrrDYNAt) | [Twitter](https://twitter.com/albumentations) | [LinkedIn](https://www.linkedin.com/company/100504475/)\n\nAlbumentations 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.\n\nHere 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:\n\n\n## Why Albumentations\n\n- 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.\n- The library provides **[a simple unified API](#a-simple-example)** to work with all data types: images (RGB-images, grayscale images, multispectral images), segmentation masks, bounding boxes, and keypoints.\n- The library contains **[more than 70 different augmentations](#list-of-augmentations)** to generate new training samples from the existing data.\n- Albumentations is **[fast](#benchmarking-results)**. We benchmark each new release to ensure that augmentations provide maximum speed.\n- 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/).\n- [**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.\n- The library is [**widely used**](#who-is-using-albumentations) in industry, deep learning research, machine learning competitions, and open source projects.\n\n## Sponsors\n\n<a href=\"https://roboflow.com/\" target=\"_blank\"><img src=\"https://avatars.githubusercontent.com/u/53104118?s=200&v=4\" width=\"100\"/></a>\n\n## Table of contents\n\n- [Albumentations](#albumentations)\n - [Why Albumentations](#why-albumentations)\n - [Sponsors](#sponsors)\n - [Table of contents](#table-of-contents)\n - [Authors](#authors)\n - [Current Maintainer](#current-maintainer)\n - [Emeritus Core Team Members](#emeritus-core-team-members)\n - [Installation](#installation)\n - [Documentation](#documentation)\n - [A simple example](#a-simple-example)\n - [Getting started](#getting-started)\n - [I am new to image augmentation](#i-am-new-to-image-augmentation)\n - [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)\n - [I want to know how to use Albumentations with deep learning frameworks](#i-want-to-know-how-to-use-albumentations-with-deep-learning-frameworks)\n - [I want to explore augmentations and see Albumentations in action](#i-want-to-explore-augmentations-and-see-albumentations-in-action)\n - [Who is using Albumentations](#who-is-using-albumentations)\n - [See also](#see-also)\n - [List of augmentations](#list-of-augmentations)\n - [Pixel-level transforms](#pixel-level-transforms)\n - [Spatial-level transforms](#spatial-level-transforms)\n - [A few more examples of **augmentations**](#a-few-more-examples-of-augmentations)\n - [Semantic segmentation on the Inria dataset](#semantic-segmentation-on-the-inria-dataset)\n - [Medical imaging](#medical-imaging)\n - [Object detection and semantic segmentation on the Mapillary Vistas dataset](#object-detection-and-semantic-segmentation-on-the-mapillary-vistas-dataset)\n - [Keypoints augmentation](#keypoints-augmentation)\n - [Benchmarking results](#benchmarking-results)\n - [System Information](#system-information)\n - [Benchmark Parameters](#benchmark-parameters)\n - [Library Versions](#library-versions)\n - [Performance Comparison](#performance-comparison)\n - [Contributing](#contributing)\n - [Community and Support](#community-and-support)\n - [Comments](#comments)\n - [Citing](#citing)\n\n## Authors\n\n### Current Maintainer\n\n[**Vladimir I. Iglovikov**](https://www.linkedin.com/in/iglovikov/) | [Kaggle Grandmaster](https://www.kaggle.com/iglovikov)\n\n### Emeritus Core Team Members\n\n[**Mikhail Druzhinin**](https://www.linkedin.com/in/mikhail-druzhinin-548229100/) | [Kaggle Expert](https://www.kaggle.com/dipetm)\n\n[**Alex Parinov**](https://www.linkedin.com/in/alex-parinov/) | [Kaggle Master](https://www.kaggle.com/creafz)\n\n[**Alexander Buslaev**](https://www.linkedin.com/in/al-buslaev/) | [Kaggle Master](https://www.kaggle.com/albuslaev)\n\n[**Eugene Khvedchenya**](https://www.linkedin.com/in/cvtalks/) | [Kaggle Grandmaster](https://www.kaggle.com/bloodaxe)\n\n## Installation\n\nAlbumentations requires Python 3.9 or higher. To install the latest version from PyPI:\n\n```bash\npip install -U albumentations\n```\n\nOther installation options are described in the [documentation](https://albumentations.ai/docs/getting_started/installation/).\n\n## Documentation\n\nThe full documentation is available at **[https://albumentations.ai/docs/](https://albumentations.ai/docs/)**.\n\n## A simple example\n\n```python\nimport albumentations as A\nimport cv2\n\n# Declare an augmentation pipeline\ntransform = A.Compose([\n A.RandomCrop(width=256, height=256),\n A.HorizontalFlip(p=0.5),\n A.RandomBrightnessContrast(p=0.2),\n])\n\n# Read an image with OpenCV and convert it to the RGB colorspace\nimage = cv2.imread(\"image.jpg\")\nimage = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)\n\n# Augment an image\ntransformed = transform(image=image)\ntransformed_image = transformed[\"image\"]\n```\n\n## Getting started\n\n### I am new to image augmentation\n\nPlease 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.\n\n### I want to use Albumentations for the specific task such as classification or segmentation\n\nIf 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.\n\n### I want to know how to use Albumentations with deep learning frameworks\n\nWe 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.\n\n### I want to explore augmentations and see Albumentations in action\n\nCheck 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).\n\n## Who is using Albumentations\n\n<a href=\"https://www.apple.com/\" target=\"_blank\"><img src=\"https://raw.githubusercontent.com/albumentations-team/albumentations.ai/main/html/assets/img/industry/apple.jpeg\" width=\"100\"/></a>\n<a href=\"https://research.google/\" target=\"_blank\"><img src=\"https://raw.githubusercontent.com/albumentations-team/albumentations.ai/main/html/assets/img/industry/google.png\" width=\"100\"/></a>\n<a href=\"https://opensource.fb.com/\" target=\"_blank\"><img src=\"https://raw.githubusercontent.com/albumentations-team/albumentations.ai/main/html/assets/img/industry/meta_research.png\" width=\"100\"/></a>\n<a href=\"https: //www.nvidia.com/en-us/research/\" target=\"_blank\"><img src=\"https://raw.githubusercontent.com/albumentations-team/albumentations.ai/main/html/assets/img/industry/nvidia_research.jpeg\" width=\"100\"/></a>\n<a href=\"https://www.amazon.science/\" target=\"_blank\"><img src=\"https://raw.githubusercontent.com/albumentations-team/albumentations.ai/main/html/assets/img/industry/amazon_science.png\" width=\"100\"/></a>\n<a href=\"https://opensource.microsoft.com/\" target=\"_blank\"><img src=\"https://raw.githubusercontent.com/albumentations-team/albumentations.ai/main/html/assets/img/industry/microsoft.png\" width=\"100\"/></a>\n<a href=\"https://engineering.salesforce.com/open-source/\" target=\"_blank\"><img src=\"https://raw.githubusercontent.com/albumentations-team/albumentations.ai/main/html/assets/img/industry/salesforce_open_source.png\" width=\"100\"/></a>\n<a href=\"https://stability.ai/\" target=\"_blank\"><img src=\"https://raw.githubusercontent.com/albumentations-team/albumentations.ai/main/html/assets/img/industry/stability.png\" width=\"100\"/></a>\n<a href=\"https://www.ibm.com/opensource/\" target=\"_blank\"><img src=\"https://raw.githubusercontent.com/albumentations-team/albumentations.ai/main/html/assets/img/industry/ibm.jpeg\" width=\"100\"/></a>\n<a href=\"https://huggingface.co/\" target=\"_blank\"><img src=\"https://raw.githubusercontent.com/albumentations-team/albumentations.ai/main/html/assets/img/industry/hugging_face.png\" width=\"100\"/></a>\n<a href=\"https://www.sony.com/en/\" target=\"_blank\"><img src=\"https://raw.githubusercontent.com/albumentations-team/albumentations.ai/main/html/assets/img/industry/sony.png\" width=\"100\"/></a>\n<a href=\"https://opensource.alibaba.com/\" target=\"_blank\"><img src=\"https://raw.githubusercontent.com/albumentations-team/albumentations.ai/main/html/assets/img/industry/alibaba.png\" width=\"100\"/></a>\n<a href=\"https://opensource.tencent.com/\" target=\"_blank\"><img src=\"https://raw.githubusercontent.com/albumentations-team/albumentations.ai/main/html/assets/img/industry/tencent.png\" width=\"100\"/></a>\n<a href=\"https://h2o.ai/\" target=\"_blank\"><img src=\"https://raw.githubusercontent.com/albumentations-team/albumentations.ai/main/html/assets/img/industry/h2o_ai.png\" width=\"100\"/></a>\n\n### See also\n\n- [A list of papers that cite Albumentations](https://scholar.google.com/citations?view_op=view_citation&citation_for_view=vkjh9X0AAAAJ:r0BpntZqJG4C).\n- [A list of teams that were using Albumentations and took high places in machine learning competitions](https://albumentations.ai/whos_using#competitions).\n- [Open source projects that use Albumentations](https://github.com/albumentations-team/albumentations/network/dependents?dependent_type=PACKAGE).\n\n## List of augmentations\n\n### Pixel-level transforms\n\nPixel-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:\n\n- [AdvancedBlur](https://explore.albumentations.ai/transform/AdvancedBlur)\n- [Blur](https://explore.albumentations.ai/transform/Blur)\n- [CLAHE](https://explore.albumentations.ai/transform/CLAHE)\n- [ChannelDropout](https://explore.albumentations.ai/transform/ChannelDropout)\n- [ChannelShuffle](https://explore.albumentations.ai/transform/ChannelShuffle)\n- [ChromaticAberration](https://explore.albumentations.ai/transform/ChromaticAberration)\n- [ColorJitter](https://explore.albumentations.ai/transform/ColorJitter)\n- [Defocus](https://explore.albumentations.ai/transform/Defocus)\n- [Downscale](https://explore.albumentations.ai/transform/Downscale)\n- [Emboss](https://explore.albumentations.ai/transform/Emboss)\n- [Equalize](https://explore.albumentations.ai/transform/Equalize)\n- [FDA](https://explore.albumentations.ai/transform/FDA)\n- [FancyPCA](https://explore.albumentations.ai/transform/FancyPCA)\n- [FromFloat](https://explore.albumentations.ai/transform/FromFloat)\n- [GaussNoise](https://explore.albumentations.ai/transform/GaussNoise)\n- [GaussianBlur](https://explore.albumentations.ai/transform/GaussianBlur)\n- [GlassBlur](https://explore.albumentations.ai/transform/GlassBlur)\n- [HistogramMatching](https://explore.albumentations.ai/transform/HistogramMatching)\n- [HueSaturationValue](https://explore.albumentations.ai/transform/HueSaturationValue)\n- [ISONoise](https://explore.albumentations.ai/transform/ISONoise)\n- [ImageCompression](https://explore.albumentations.ai/transform/ImageCompression)\n- [InvertImg](https://explore.albumentations.ai/transform/InvertImg)\n- [MedianBlur](https://explore.albumentations.ai/transform/MedianBlur)\n- [MotionBlur](https://explore.albumentations.ai/transform/MotionBlur)\n- [MultiplicativeNoise](https://explore.albumentations.ai/transform/MultiplicativeNoise)\n- [Normalize](https://explore.albumentations.ai/transform/Normalize)\n- [PixelDistributionAdaptation](https://explore.albumentations.ai/transform/PixelDistributionAdaptation)\n- [PlanckianJitter](https://explore.albumentations.ai/transform/PlanckianJitter)\n- [Posterize](https://explore.albumentations.ai/transform/Posterize)\n- [RGBShift](https://explore.albumentations.ai/transform/RGBShift)\n- [RandomBrightnessContrast](https://explore.albumentations.ai/transform/RandomBrightnessContrast)\n- [RandomFog](https://explore.albumentations.ai/transform/RandomFog)\n- [RandomGamma](https://explore.albumentations.ai/transform/RandomGamma)\n- [RandomGravel](https://explore.albumentations.ai/transform/RandomGravel)\n- [RandomRain](https://explore.albumentations.ai/transform/RandomRain)\n- [RandomShadow](https://explore.albumentations.ai/transform/RandomShadow)\n- [RandomSnow](https://explore.albumentations.ai/transform/RandomSnow)\n- [RandomSunFlare](https://explore.albumentations.ai/transform/RandomSunFlare)\n- [RandomToneCurve](https://explore.albumentations.ai/transform/RandomToneCurve)\n- [RingingOvershoot](https://explore.albumentations.ai/transform/RingingOvershoot)\n- [Sharpen](https://explore.albumentations.ai/transform/Sharpen)\n- [Solarize](https://explore.albumentations.ai/transform/Solarize)\n- [Spatter](https://explore.albumentations.ai/transform/Spatter)\n- [Superpixels](https://explore.albumentations.ai/transform/Superpixels)\n- [TemplateTransform](https://explore.albumentations.ai/transform/TemplateTransform)\n- [TextImage](https://explore.albumentations.ai/transform/TextImage)\n- [ToFloat](https://explore.albumentations.ai/transform/ToFloat)\n- [ToGray](https://explore.albumentations.ai/transform/ToGray)\n- [ToRGB](https://explore.albumentations.ai/transform/ToRGB)\n- [ToSepia](https://explore.albumentations.ai/transform/ToSepia)\n- [UnsharpMask](https://explore.albumentations.ai/transform/UnsharpMask)\n- [ZoomBlur](https://explore.albumentations.ai/transform/ZoomBlur)\n\n### Spatial-level transforms\n\nSpatial-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.\n\n| Transform | Image | Mask | BBoxes | Keypoints |\n| ------------------------------------------------------------------------------------------------ | :---: | :--: | :----: | :-------: |\n| [Affine](https://explore.albumentations.ai/transform/Affine) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [BBoxSafeRandomCrop](https://explore.albumentations.ai/transform/BBoxSafeRandomCrop) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [CenterCrop](https://explore.albumentations.ai/transform/CenterCrop) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [CoarseDropout](https://explore.albumentations.ai/transform/CoarseDropout) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [Crop](https://explore.albumentations.ai/transform/Crop) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [CropAndPad](https://explore.albumentations.ai/transform/CropAndPad) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [CropNonEmptyMaskIfExists](https://explore.albumentations.ai/transform/CropNonEmptyMaskIfExists) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [D4](https://explore.albumentations.ai/transform/D4) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [ElasticTransform](https://explore.albumentations.ai/transform/ElasticTransform) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [GridDistortion](https://explore.albumentations.ai/transform/GridDistortion) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [GridDropout](https://explore.albumentations.ai/transform/GridDropout) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [GridElasticDeform](https://explore.albumentations.ai/transform/GridElasticDeform) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [HorizontalFlip](https://explore.albumentations.ai/transform/HorizontalFlip) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [Lambda](https://explore.albumentations.ai/transform/Lambda) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [LongestMaxSize](https://explore.albumentations.ai/transform/LongestMaxSize) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [MaskDropout](https://explore.albumentations.ai/transform/MaskDropout) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [Morphological](https://explore.albumentations.ai/transform/Morphological) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [NoOp](https://explore.albumentations.ai/transform/NoOp) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [OpticalDistortion](https://explore.albumentations.ai/transform/OpticalDistortion) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [OverlayElements](https://explore.albumentations.ai/transform/OverlayElements) | \u2713 | \u2713 | | |\n| [PadIfNeeded](https://explore.albumentations.ai/transform/PadIfNeeded) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [Perspective](https://explore.albumentations.ai/transform/Perspective) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [PiecewiseAffine](https://explore.albumentations.ai/transform/PiecewiseAffine) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [PixelDropout](https://explore.albumentations.ai/transform/PixelDropout) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [RandomCrop](https://explore.albumentations.ai/transform/RandomCrop) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [RandomCropFromBorders](https://explore.albumentations.ai/transform/RandomCropFromBorders) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [RandomGridShuffle](https://explore.albumentations.ai/transform/RandomGridShuffle) | \u2713 | \u2713 | | \u2713 |\n| [RandomResizedCrop](https://explore.albumentations.ai/transform/RandomResizedCrop) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [RandomRotate90](https://explore.albumentations.ai/transform/RandomRotate90) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [RandomScale](https://explore.albumentations.ai/transform/RandomScale) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [RandomSizedBBoxSafeCrop](https://explore.albumentations.ai/transform/RandomSizedBBoxSafeCrop) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [RandomSizedCrop](https://explore.albumentations.ai/transform/RandomSizedCrop) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [Resize](https://explore.albumentations.ai/transform/Resize) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [Rotate](https://explore.albumentations.ai/transform/Rotate) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [SafeRotate](https://explore.albumentations.ai/transform/SafeRotate) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [ShiftScaleRotate](https://explore.albumentations.ai/transform/ShiftScaleRotate) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [SmallestMaxSize](https://explore.albumentations.ai/transform/SmallestMaxSize) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [Transpose](https://explore.albumentations.ai/transform/Transpose) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [VerticalFlip](https://explore.albumentations.ai/transform/VerticalFlip) | \u2713 | \u2713 | \u2713 | \u2713 |\n| [XYMasking](https://explore.albumentations.ai/transform/XYMasking) | \u2713 | \u2713 | \u2713 | \u2713 |\n\n## A few more examples of **augmentations**\n\n### Semantic segmentation on the Inria dataset\n\n\n\n### Medical imaging\n\n\n\n### Object detection and semantic segmentation on the Mapillary Vistas dataset\n\n\n\n### Keypoints augmentation\n\n<img src=\"https://habrastorage.org/webt/e-/6k/z-/e-6kz-fugp2heak3jzns3bc-r8o.jpeg\" width=100%>\n\n## Benchmarking results\n\n### System Information\n\n- Platform: macOS-15.0.1-arm64-arm-64bit\n- Processor: arm\n- CPU Count: 10\n- Python Version: 3.12.7\n\n### Benchmark Parameters\n\n- Number of images: 1000\n- Runs per transform: 10\n- Max warmup iterations: 1000\n\n### Library Versions\n\n- albumentations: 1.4.20\n- augly: 1.0.0\n- imgaug: 0.4.0\n- kornia: 0.7.3\n- torchvision: 0.20.0\n\n## Performance Comparison\n\n| Transform | albumentations<br>1.4.20 | augly<br>1.0.0 | imgaug<br>0.4.0 | kornia<br>0.7.3 | torchvision<br>0.20.0 |\n|:------------------|:---------------------------|:-----------------|:------------------|:------------------|:------------------------|\n| HorizontalFlip | **8618 \u00b1 1233** | 4807 \u00b1 818 | 6042 \u00b1 788 | 390 \u00b1 106 | 914 \u00b1 67 |\n| VerticalFlip | **22847 \u00b1 2031** | 9153 \u00b1 1291 | 10931 \u00b1 1844 | 1212 \u00b1 402 | 3198 \u00b1 200 |\n| Rotate | **1146 \u00b1 79** | 1119 \u00b1 41 | 1136 \u00b1 218 | 143 \u00b1 11 | 181 \u00b1 11 |\n| Affine | 682 \u00b1 192 | - | **774 \u00b1 97** | 147 \u00b1 9 | 130 \u00b1 12 |\n| Equalize | **892 \u00b1 61** | - | 581 \u00b1 54 | 152 \u00b1 19 | 479 \u00b1 12 |\n| RandomCrop80 | **47341 \u00b1 20523** | 25272 \u00b1 1822 | 11503 \u00b1 441 | 1510 \u00b1 230 | 32109 \u00b1 1241 |\n| ShiftRGB | **2349 \u00b1 76** | - | 1582 \u00b1 65 | - | - |\n| Resize | **2316 \u00b1 166** | 611 \u00b1 78 | 1806 \u00b1 63 | 232 \u00b1 24 | 195 \u00b1 4 |\n| RandomGamma | **8675 \u00b1 274** | - | 2318 \u00b1 269 | 108 \u00b1 13 | - |\n| Grayscale | **3056 \u00b1 47** | 2720 \u00b1 932 | 1681 \u00b1 156 | 289 \u00b1 75 | 1838 \u00b1 130 |\n| RandomPerspective | 412 \u00b1 38 | - | **554 \u00b1 22** | 86 \u00b1 11 | 96 \u00b1 5 |\n| GaussianBlur | **1728 \u00b1 89** | 242 \u00b1 4 | 1090 \u00b1 65 | 176 \u00b1 18 | 79 \u00b1 3 |\n| MedianBlur | **868 \u00b1 60** | - | 813 \u00b1 30 | 5 \u00b1 0 | - |\n| MotionBlur | **4047 \u00b1 67** | - | 612 \u00b1 18 | 73 \u00b1 2 | - |\n| Posterize | **9094 \u00b1 301** | - | 2097 \u00b1 68 | 430 \u00b1 49 | 3196 \u00b1 185 |\n| JpegCompression | **918 \u00b1 23** | 778 \u00b1 5 | 459 \u00b1 35 | 71 \u00b1 3 | 625 \u00b1 17 |\n| GaussianNoise | 166 \u00b1 12 | 67 \u00b1 2 | **206 \u00b1 11** | 75 \u00b1 1 | - |\n| Elastic | 201 \u00b1 5 | - | **235 \u00b1 20** | 1 \u00b1 0 | 2 \u00b1 0 |\n| Clahe | **454 \u00b1 22** | - | 335 \u00b1 43 | 94 \u00b1 9 | - |\n| CoarseDropout | **13368 \u00b1 744** | - | 671 \u00b1 38 | 536 \u00b1 87 | - |\n| Blur | **5267 \u00b1 543** | 246 \u00b1 3 | 3807 \u00b1 325 | - | - |\n| ColorJitter | **628 \u00b1 55** | 255 \u00b1 13 | - | 55 \u00b1 18 | 46 \u00b1 2 |\n| Brightness | **8956 \u00b1 300** | 1163 \u00b1 86 | - | 472 \u00b1 101 | 429 \u00b1 20 |\n| Contrast | **8879 \u00b1 1426** | 736 \u00b1 79 | - | 425 \u00b1 52 | 335 \u00b1 35 |\n| RandomResizedCrop | **2828 \u00b1 186** | - | - | 287 \u00b1 58 | 511 \u00b1 10 |\n| Normalize | **1196 \u00b1 56** | - | - | 626 \u00b1 40 | 519 \u00b1 12 |\n| PlankianJitter | **2204 \u00b1 385** | - | - | 813 \u00b1 211 | - |\n\n## Contributing\n\nTo create a pull request to the repository, follow the documentation at [CONTRIBUTING.md](CONTRIBUTING.md)\n\n\n\n## Community and Support\n\n- [Twitter](https://twitter.com/albumentations)\n- [Discord](https://discord.gg/AKPrrDYNAt)\n- [LinkedIn](https://www.linkedin.com/company/albumentations/)\n\n## Comments\n\nIn 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)\n\n```python\ncv2.setNumThreads(0)\ncv2.ocl.setUseOpenCL(False)\n```\n\n## Citing\n\nIf 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):\n\n```bibtex\n@Article{info11020125,\n AUTHOR = {Buslaev, Alexander and Iglovikov, Vladimir I. and Khvedchenya, Eugene and Parinov, Alex and Druzhinin, Mikhail and Kalinin, Alexandr A.},\n TITLE = {Albumentations: Fast and Flexible Image Augmentations},\n JOURNAL = {Information},\n VOLUME = {11},\n YEAR = {2020},\n NUMBER = {2},\n ARTICLE-NUMBER = {125},\n URL = {https://www.mdpi.com/2078-2489/11/2/125},\n ISSN = {2078-2489},\n DOI = {10.3390/info11020125}\n}\n```\n",
"bugtrack_url": null,
"license": "MIT License Copyright (c) 2017 Vladimir Iglovikov, Alexander Buslaev, Alexander Parinov, Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the \"Software\"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. ",
"summary": "Fast, flexible, and advanced image augmentation library for deep learning and computer vision. Albumentations offers a wide range of transformations for images, masks, bounding boxes, and keypoints, with optimized performance and seamless integration into ML workflows.",
"version": "1.4.21",
"project_urls": {
"Homepage": "https://albumentations.ai"
},
"split_keywords": [
"artificial intelligence",
" bounding box",
" computer vision",
" computer vision library",
" data augmentation",
" data preprocessing",
" deep learning",
" fast augmentation",
" image augmentation",
" image classification",
" image processing",
" image transformation",
" instance segmentation",
" keras",
" keypoint detection",
" machine learning",
" object detection",
" pytorch",
" semantic segmentation",
" tensorflow"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "b0e5466001d1f1567932d5ccd499287e31508f405d113fe97f961312d9a64597",
"md5": "09b74d3aee5ce0d1001bff88cb987d3f",
"sha256": "5e346b39df5f782cebcccee320fee5bcf816243e560ba17eb44d8dc3825339b4"
},
"downloads": -1,
"filename": "albumentations-1.4.21-py3-none-any.whl",
"has_sig": false,
"md5_digest": "09b74d3aee5ce0d1001bff88cb987d3f",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.9",
"size": 227933,
"upload_time": "2024-11-01T00:16:00",
"upload_time_iso_8601": "2024-11-01T00:16:00.912786Z",
"url": "https://files.pythonhosted.org/packages/b0/e5/466001d1f1567932d5ccd499287e31508f405d113fe97f961312d9a64597/albumentations-1.4.21-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "ac7aab6f36f33cfb7639b8a1df4656624a94b03a164c0a063747cb0c59c8fdfa",
"md5": "fa133c4df532e666025ad2fb7eb1b70c",
"sha256": "62717725a9a046acdd65ea82a25e5e46e37349e2e5073fbf5a1d99f49cb49601"
},
"downloads": -1,
"filename": "albumentations-1.4.21.tar.gz",
"has_sig": false,
"md5_digest": "fa133c4df532e666025ad2fb7eb1b70c",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.9",
"size": 211612,
"upload_time": "2024-11-01T00:16:02",
"upload_time_iso_8601": "2024-11-01T00:16:02.760318Z",
"url": "https://files.pythonhosted.org/packages/ac/7a/ab6f36f33cfb7639b8a1df4656624a94b03a164c0a063747cb0c59c8fdfa/albumentations-1.4.21.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-11-01 00:16:02",
"github": false,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"lcname": "albumentations"
}
