Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/lewis-morris/image_augment
Image augmentation for CNN's. Aiming to increase generalization.
https://github.com/lewis-morris/image_augment
augmentation image numpy python
Last synced: 16 days ago
JSON representation
Image augmentation for CNN's. Aiming to increase generalization.
- Host: GitHub
- URL: https://github.com/lewis-morris/image_augment
- Owner: lewis-morris
- License: gpl-3.0
- Created: 2020-04-27T04:16:06.000Z (over 4 years ago)
- Default Branch: master
- Last Pushed: 2020-04-28T11:29:45.000Z (over 4 years ago)
- Last Synced: 2024-10-16T20:12:07.267Z (about 1 month ago)
- Topics: augmentation, image, numpy, python
- Language: Jupyter Notebook
- Homepage:
- Size: 40.9 MB
- Stars: 0
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Basic image augmenter
Basic image augmentation for CNN's. Aiming to add regularisation to your model at the input layer.Can be used in a pytorch pipeline
## Setup
Use pip to install
`pip install image-augment`
[PyPI](https://pypi.org/project/image-augment/)
## Features
Cutout - blocking out random segments of the image - details on this technique can be found here. https://arxiv.org/pdf/1708.04552.pdf
Noise - randomly adds noise to image
Mirroring - randomly flips image on vertical and horizontal axis.
Rotation - randomly rotates image
## Basic Usage.
```python
from ImageAugment import basic
from PIL import Imageimg = Image.open(./example_images/messi5.jpg)
img
```![messi](./example_images/messi5.jpg "Messi")
```python
new_image = basic()(img)
new_image
```![messi](./example_images/messi_changed.jpg "Messi")
# For more info check the example jupyter notebook
[Jupyter examples](./example.ipynb)