Image augmentation library in Python for machine learning.

Overview

AugmentorLogo

Augmentor is an image augmentation library in Python for machine learning. It aims to be a standalone library that is platform and framework independent, which is more convenient, allows for finer grained control over augmentation, and implements the most real-world relevant augmentation techniques. It employs a stochastic approach using building blocks that allow for operations to be pieced together in a pipeline.

PyPI Supported Python Versions Documentation Status Build Status License Project Status: Active – The project has reached a stable, usable state and is being actively developed. Binder

Installation

Augmentor is written in Python. A Julia version of the package is also being developed as a sister project and is available here.

Install using pip from the command line:

pip install Augmentor

See the documentation for building from source. To upgrade from a previous version, use pip install Augmentor --upgrade.

Documentation

Complete documentation can be found on Read the Docs: http://augmentor.readthedocs.io/

Quick Start Guide and Usage

The purpose of Augmentor is to automate image augmentation (artificial data generation) in order to expand datasets as input for machine learning algorithms, especially neural networks and deep learning.

The package works by building an augmentation pipeline where you define a series of operations to perform on a set of images. Operations, such as rotations or transforms, are added one by one to create an augmentation pipeline: when complete, the pipeline can be executed and an augmented dataset is created.

To begin, instantiate a Pipeline object that points to a directory on your file system:

import Augmentor
p = Augmentor.Pipeline("/path/to/images")

You can then add operations to the Pipeline object p as follows:

p.rotate(probability=0.7, max_left_rotation=10, max_right_rotation=10)
p.zoom(probability=0.5, min_factor=1.1, max_factor=1.5)

Every function requires you to specify a probability, which is used to decide if an operation is applied to an image as it is passed through the augmentation pipeline.

Once you have created a pipeline, you can sample from it like so:

p.sample(10000)

which will generate 10,000 augmented images based on your specifications. By default these will be written to the disk in a directory named output relative to the path specified when initialising the p pipeline object above.

If you wish to process each image in the pipeline exactly once, use process():

p.process()

This function might be useful for resizing a dataset for example. It would make sense to create a pipeline where all of its operations have their probability set to 1 when using the process() method.

Multi-threading

Augmentor (version >=0.2.1) now uses multi-threading to increase the speed of generating images.

This may slow down some pipelines if the original images are very small. Set multi_threaded to False if slowdown is experienced:

p.sample(100, multi_threaded=False)

However, by default the sample() function uses multi-threading. This is currently only implemented when saving to disk. Generators will use multi-threading in the next version update.

Ground Truth Data

Images can be passed through the pipeline in groups of two or more so that ground truth data can be identically augmented.

Original image and mask[3] Augmented original and mask images
OriginalMask AugmentedMask

To augment ground truth data in parallel to any original data, add a ground truth directory to a pipeline using the ground_truth() function:

p = Augmentor.Pipeline("/path/to/images")
# Point to a directory containing ground truth data.
# Images with the same file names will be added as ground truth data
# and augmented in parallel to the original data.
p.ground_truth("/path/to/ground_truth_images")
# Add operations to the pipeline as normal:
p.rotate(probability=1, max_left_rotation=5, max_right_rotation=5)
p.flip_left_right(probability=0.5)
p.zoom_random(probability=0.5, percentage_area=0.8)
p.flip_top_bottom(probability=0.5)
p.sample(50)

Multiple Mask/Image Augmentation

Using the DataPipeline class (Augmentor version >= 0.2.3), images that have multiple associated masks can be augmented:

Multiple Mask Augmentation
MultipleMask

Arbitrarily long lists of images can be passed through the pipeline in groups and augmented identically using the DataPipeline class. This is useful for ground truth images that have several masks, for example.

In the example below, the images and their masks are contained in the images data structure (as lists of lists), while their labels are contained in y:

p = Augmentor.DataPipeline(images, y)
p.rotate(1, max_left_rotation=5, max_right_rotation=5)
p.flip_top_bottom(0.5)
p.zoom_random(1, percentage_area=0.5)

augmented_images, labels = p.sample(100)

The DataPipeline returns images directly (augmented_images above), and does not save them to disk, nor does it read data from the disk. Images are passed directly to DataPipeline during initialisation.

For details of the images data structure and how to create it, see the Multiple-Mask-Augmentation.ipynb Jupyter notebook.

Generators for Keras and PyTorch

If you do not wish to save to disk, you can use a generator (in this case with Keras):

g = p.keras_generator(batch_size=128)
images, labels = next(g)

which returns a batch of images of size 128 and their corresponding labels. Generators return data indefinitely, and can be used to train neural networks with augmented data on the fly.

Alternatively, you can integrate it with PyTorch:

import torchvision
transforms = torchvision.transforms.Compose([
    p.torch_transform(),
    torchvision.transforms.ToTensor(),
])

Main Features

Elastic Distortions

Using elastic distortions, one image can be used to generate many images that are real-world feasible and label preserving:

Input Image Augmented Images
eight_hand_drawn_border eights_border

The input image has a 1 pixel black border to emphasise that you are getting distortions without changing the size or aspect ratio of the original image, and without any black/transparent padding around the newly generated images.

The functionality can be more clearly seen here:

Original Image[1] Random distortions applied
Original Distorted

Perspective Transforms

There are a total of 12 different types of perspective transform available. Four of the most common are shown below.

Tilt Left Tilt Right Tilt Forward Tilt Backward
TiltLeft Original Original Original

The remaining eight types of transform are as follows:

Skew Type 0 Skew Type 1 Skew Type 2 Skew Type 3
Skew0 Skew1 Skew2 Skew3
Skew Type 4 Skew Type 5 Skew Type 6 Skew Type 7
Skew4 Skew5 Skew6 Skew7

Size Preserving Rotations

Rotations by default preserve the file size of the original images:

Original Image Rotated 10 degrees, automatically cropped
Original Rotate

Compared to rotations by other software:

Original Image Rotated 10 degrees
Original Rotate

Size Preserving Shearing

Shearing will also automatically crop the correct area from the sheared image, so that you have an image with no black space or padding.

Original image Shear (x-axis) 20 degrees Shear (y-axis) 20 degrees
Original ShearX ShearY

Compare this to how this is normally done:

Original image Shear (x-axis) 20 degrees Shear (y-axis) 20 degrees
Original ShearX ShearY

Cropping

Cropping can also be handled in a manner more suitable for machine learning image augmentation:

Original image Random crops + resize operation
Original Original

Random Erasing

Random Erasing is a technique used to make models robust to occlusion. This may be useful for training neural networks used in object detection in navigation scenarios, for example.

Original image[2] Random Erasing
Original Original

See the Pipeline.random_erasing() documentation for usage.

Chaining Operations in a Pipeline

With only a few operations, a single image can be augmented to produce large numbers of new, label-preserving samples:

Original image Distortions + mirroring
Original DistortFlipFlop

In the example above, we have applied three operations: first we randomly distort the image, then we flip it horizontally with a probability of 0.5 and then vertically with a probability of 0.5. We then sample from this pipeline 100 times to create 100 new data.

p.random_distortion(probability=1, grid_width=4, grid_height=4, magnitude=8)
p.flip_left_right(probability=0.5)
p.flip_top_bottom(probability=0.5)
p.sample(100)

Tutorial Notebooks

Integration with Keras using Generators

Augmentor can be used as a replacement for Keras' augmentation functionality. Augmentor can create a generator which produces augmented data indefinitely, according to the pipeline you have defined. See the following notebooks for details:

  • Reading images from a local directory, augmenting them at run-time, and using a generator to pass the augmented stream of images to a Keras convolutional neural network, see Augmentor_Keras.ipynb
  • Augmenting data in-memory (in array format) and using a generator to pass these new images to the Keras neural network, see Augmentor_Keras_Array_Data.ipynb

Per-Class Augmentation Strategies

Augmentor allows for pipelines to be defined per class. That is, you can define different augmentation strategies on a class-by-class basis for a given classification problem.

See an example of this in the following Jupyter notebook: Per_Class_Augmentation_Strategy.ipynb

Complete Example

Let's perform an augmentation task on a single image, demonstrating the pipeline and several features of Augmentor.

First import the package and initialise a Pipeline object by pointing it to a directory containing your images:

import Augmentor

p = Augmentor.Pipeline("/home/user/augmentor_data_tests")

Now you can begin adding operations to the pipeline object:

p.rotate90(probability=0.5)
p.rotate270(probability=0.5)
p.flip_left_right(probability=0.8)
p.flip_top_bottom(probability=0.3)
p.crop_random(probability=1, percentage_area=0.5)
p.resize(probability=1.0, width=120, height=120)

Once you have added the operations you require, you can sample images from this pipeline:

p.sample(100)

Some sample output:

Input Image[3] Augmented Images
Original Augmented

The augmented images may be useful for a boundary detection task, for example.

Licence and Acknowledgements

Augmentor is made available under the terms of the MIT Licence. See Licence.md.

[1] Checkerboard image obtained from Wikimedia Commons and is in the public domain: https://commons.wikimedia.org/wiki/File:Checkerboard_pattern.svg

[2] Street view image is in the public domain: http://stokpic.com/project/italian-city-street-with-shoppers/

[3] Skin lesion image obtained from the ISIC Archive:

You can use urllib to obtain the skin lesion image in order to reproduce the augmented images above:

>>> from urllib import urlretrieve
>>> im_url = "https://isic-archive.com:443/api/v1/image/5436e3abbae478396759f0cf/download"
>>> urlretrieve(im_url, "ISIC_0000000.jpg")
('ISIC_0000000.jpg', <httplib.HTTPMessage instance at 0x7f7bd949a950>)

Note: For Python 3, use from urllib.request import urlretrieve.

Logo created at LogoMakr.com

Tests

To run the automated tests, clone the repository and run:

$ py.test -v

from the command line. To view the CI tests that are run after each commit, see https://travis-ci.org/mdbloice/Augmentor.

Citing Augmentor

If you find this package useful and wish to cite it, you can use

Marcus D Bloice, Peter M Roth, Andreas Holzinger, Biomedical image augmentation using Augmentor, Bioinformatics, https://doi.org/10.1093/bioinformatics/btz259

Asciicast

Click the preview below to view a video demonstration of Augmentor in use:

asciicast

Owner
Marcus D. Bloice
Researcher in applied machine learning for healthcare, Medical University of Graz, Austria.
Marcus D. Bloice
Building a real-time environment using webcam frame division in OpenCV and classify cropped images using a fine-tuned vision transformers on hybryd datasets samples for facial emotion recognition.

Visual Transformer for Facial Emotion Recognition (FER) This project has the aim to build an efficient Visual Transformer for the Facial Emotion Recog

Mario Sessa 8 Dec 12, 2022
LVI-SAM: Tightly-coupled Lidar-Visual-Inertial Odometry via Smoothing and Mapping

LVI-SAM This repository contains code for a lidar-visual-inertial odometry and mapping system, which combines the advantages of LIO-SAM and Vins-Mono

Tixiao Shan 1.1k Dec 27, 2022
Tensorflow implementation of Character-Aware Neural Language Models.

Character-Aware Neural Language Models Tensorflow implementation of Character-Aware Neural Language Models. The original code of author can be found h

Taehoon Kim 751 Dec 26, 2022
Software associated to AAAI paper "Planning with Biological Neurons and Synapses"

jBrain Software associated with the AAAI 2022 paper Francesco D'Amore, Daniel Mitropolsky, Pierluigi Crescenzi, Emanuele Natale, Christos H. Papadimit

Pierluigi Crescenzi 1 Apr 10, 2022
Code for Robust Contrastive Learning against Noisy Views

Robust Contrastive Learning against Noisy Views This repository provides a PyTorch implementation of the Robust InfoNCE loss proposed in paper Robust

Ching-Yao Chuang 53 Jan 08, 2023
Code for SentiBERT: A Transferable Transformer-Based Architecture for Compositional Sentiment Semantics (ACL'2020).

SentiBERT Code for SentiBERT: A Transferable Transformer-Based Architecture for Compositional Sentiment Semantics (ACL'2020). https://arxiv.org/abs/20

Da Yin 66 Aug 13, 2022
Clean and readable code for Decision Transformer: Reinforcement Learning via Sequence Modeling

Minimal implementation of Decision Transformer: Reinforcement Learning via Sequence Modeling in PyTorch for mujoco control tasks in OpenAI gym

Nikhil Barhate 104 Jan 06, 2023
Analysis code and Latex source of the manuscript describing the conditional permutation test of confounding bias in predictive modelling.

Git repositoty of the manuscript entitled Statistical quantification of confounding bias in predictive modelling by Tamas Spisak The manuscript descri

PNI - Predictive Neuroimaging Lab, University Hospital Essen, Germany 0 Nov 22, 2021
This library provides an abstraction to perform Model Versioning using Weight & Biases.

Description This library provides an abstraction to perform Model Versioning using Weight & Biases. Features Version a new trained model Promote a mod

Hector Lopez Almazan 2 Jan 28, 2022
BMVC 2021 Oral: code for BI-GCN: Boundary-Aware Input-Dependent Graph Convolution for Biomedical Image Segmentation

BMVC 2021 BI-GConv: Boundary-Aware Input-Dependent Graph Convolution for Biomedical Image Segmentation Necassary Dependencies: PyTorch 1.2.0 Python 3.

Yanda Meng 15 Nov 08, 2022
Unified unsupervised and semi-supervised domain adaptation network for cross-scenario face anti-spoofing, Pattern Recognition

USDAN The implementation of Unified unsupervised and semi-supervised domain adaptation network for cross-scenario face anti-spoofing, which is accepte

11 Nov 03, 2022
Real Time Object Detection and Classification using Yolo Algorithm.

Real time Object detection & Classification using YOLO algorithm. Real Time Object Detection and Classification using Yolo Algorithm. What is Object D

Ketan Chawla 1 Apr 17, 2022
Capstone-Project-2 - A game program written in the Python language

Capstone-Project-2 My Pygame Game Information: Description This Pygame project i

Nhlakanipho Khulekani Hlophe 1 Jan 04, 2022
Hierarchical Few-Shot Generative Models

Hierarchical Few-Shot Generative Models Giorgio Giannone, Ole Winther This repo contains code and experiments for the paper Hierarchical Few-Shot Gene

Giorgio Giannone 6 Dec 12, 2022
Using Machine Learning to Create High-Res Fine Art

BIG.art: Using Machine Learning to Create High-Res Fine Art How to use GLIDE and BSRGAN to create ultra-high-resolution paintings with fine details By

Robert A. Gonsalves 13 Nov 27, 2022
Codes for the ICCV'21 paper "FREE: Feature Refinement for Generalized Zero-Shot Learning"

FREE This repository contains the reference code for the paper "FREE: Feature Refinement for Generalized Zero-Shot Learning". [arXiv][Paper] 1. Prepar

Shiming Chen 28 Jul 29, 2022
Official implementation of "Generating 3D Molecules for Target Protein Binding"

Generating 3D Molecules for Target Protein Binding This is the official implementation of the GraphBP method proposed in the following paper. Meng Liu

DIVE Lab, Texas A&M University 74 Dec 07, 2022
This is the official PyTorch implementation of the paper "TransFG: A Transformer Architecture for Fine-grained Recognition" (Ju He, Jie-Neng Chen, Shuai Liu, Adam Kortylewski, Cheng Yang, Yutong Bai, Changhu Wang, Alan Yuille).

TransFG: A Transformer Architecture for Fine-grained Recognition Official PyTorch code for the paper: TransFG: A Transformer Architecture for Fine-gra

Ju He 307 Jan 03, 2023
A Home Assistant custom component for Lobe. Lobe is an AI tool that can classify images.

Lobe This is a Home Assistant custom component for Lobe. Lobe is an AI tool that can classify images. This component lets you easily use an exported m

Kendell R 4 Feb 28, 2022
Voice control for Garry's Mod

WIP: Talonvoice GMod integrations Very work in progress voice control demo for Garry's Mod. HOWTO Install https://talonvoice.com/ Press https://i.imgu

Meta Construct 5 Nov 15, 2022