Image Augmentation Pyhton Library

tags: README

Getting Started

This Repository can help us to facilitate image data augment and we can deploy this lib on any computer vision task. If you want to train a deep learning model for image classification or object detection data augmentation will make your model more robust.

Requirements

• Python
• Open CV
• Numpy

Usage

First, git this repo to local

$ git clone https://github.com/allen108108/Image_Augmentation.git
$ cd Image_Augmentation

src is the images source folder, and this is necessary parameter.In the source images, you can give a propotion -f to dicide how much images you want to augment.

Finally, check what augmentation type that you want : (Note : -hf, -vf, -cs, -hs or vs that you must choose only one type. )

$ python run.py
  -h, --help  show this help message and exit
   src        image src path
  -f F        proportion of augmentation 
  -hf         Horizontal Flip
  -vf         Vertical Flip
  -cs         Color Shift
  -hs         Horizontal Shift
  -vs         Vertical Shift

The augmented images saved in ./augmention/<augment type>. You can change the location in run.py.