Albumentations: fast and flexible image augmentations

By Alexander V. Buslaev, Alex Parinov, Eugene Khvedchenya, Vladimir I. Iglovikov, Alexandr A Kalinin. 2018

2,644 citations recorded by the source corpus.

Data augmentation is a commonly used technique for increasing both the size and the diversity of labeled training sets by leveraging input transformations that preserve corresponding output labels. In computer vision, image augmentations have become a common implicit regularization technique to combat overfitting in deep learning models and are ubiquitously used to improve performance. While most deep learning frameworks implement basic image transformations, the list is typically limited to some variations of flipping, rotating, scaling, and cropping. Moreover, image processing speed varies in existing image augmentation libraries. We present Albumentations, a fast and flexible open source library for image augmentation with many various image transform operations available that is also an easy-to-use wrapper around other augmentation libraries. We discuss the design principles that drove the implementation of Albumentations and give an overview of the key features and distinct capabilities. Finally, we provide examples of image augmentations for different computer vision tasks and demonstrate that Albumentations is faster than other commonly used image augmentation tools on most image transform operations.

Subfields: Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

Citations in the corpus

Breast Tumor Cellularity Assessment using Deep Neural Networks (2019, methodology): “During training, we dynamically augmented images with vertical and horizontal flips, rotation, gamma, hue, and saturation utilizing the Albumentations library [43].”
TorchIO: A Python library for efficient loading, preprocessing, augmentation and patch-based sampling of medical images in deep learning (2020, methodology): “Multiple libraries with support for computer vision data augmentation have appeared in the last years, such as Albumentations [Buslaev et al., 2020], Augmentor [Bloice et al.”
Benchmarking Low-Light Image Enhancement and Beyond (2021, methodology): “Data Augmentation We adopt RandomCrop, HorizontalFlip, RandomSizedBBoxSafeCrop from the Albumentation Library (Buslaev et al. 2018) to prevent over-ﬁtting and construct a more robust model.”
Paint by Example: Exemplar-based Image Editing with Diffusion Models (2022, positive): “3) from Albumentations [1] for image augmentation.”
CellViT: Vision Transformers for Precise Cell Segmentation and Classification (2023, methodology): “To augment images and masks, we used the Albumen-tations library [71].”
Transformation-Consistent Self-Ensembling Model for Semisupervised Medical Image Segmentation (2019, background): “Moreover, recent automatic augmentation search works [66– 68] explored the best transformations for a specific dataset.”
CLoDSA: a tool for augmentation in classification, localization, detection, semantic segmentation and instance segmentation tasks (2019, background): “Imgaug [11] provides more than 40 augmentation techniques, and albumentations [47] is the fastest augmentation library.”
AugLy: Data Augmentations for Robustness (2022, methodology): “A majority of libraries focus on images [1, 19, 14] and text [18, 8, 21], however audio [20, 29, 15] and video [32, 6, 16] augmentation libraries do exist as well with more limited augmentations (see Section 4 for in-depth comparisons between AugLy and existing libraries for each modality).”
DeblurGAN-v2: Deblurring (Orders-of-Magnitude) Faster and Better (2019, methodology): “We then use the albumentations library [4] to further add Gaussian and speckle Noise, JPEG compression, and up-scaling artifacts to those images.”
Synthetic Data for Deep Learning (2019, methodology): “These days, data augmentation is a crucial part of virtually every computer vision pipeline; we refer to the surveys [544,630] and especially recommend the Albumentations library [78] that has proven invaluable in our practice, but in this survey we concentrate on synthetic data rather than augmenta”
Semi-Supervised Segmentation of Salt Bodies in Seismic Images using an Ensemble of Convolutional Neural Networks (2019, methodology): “The first idea is to use Test Time Augmentations (TTA): instead of predicting on a single test image, we average predictions on the original test image and its horizontal flip [6].”
Detecting Deepfakes with Self-Blended Images (2022, methodology): “For the color and frequency transformations, we adopt RGBShift, HueSaturationValue, RandomBrightnessContrast, Downscale, and Sharpen from a widely used image processing toolbox [11].”