Document Image Augmentation

Introduction

Document Image Augmentation is simple to use tool for performing Augmentation on Document-based Images.
This tool provides a total of 4 types of document Augmentation

Following are Releasing Soon

1. Add support for Ground Truth
2. Add Flip and Rotate
3. Add Another Document Specific Augmentation
4. Modify Color Changer to automatically detect background color

Quick Start

1. Requirements: numpy & opencv-python

• pip install numpy
• pip install opencv-python

2. Run DocAug.py; examples:

• Process a single image:
  • python DocAug.py --input_image_filename ../images/image1.jpg
• Process all images in a directory :
  • python DocAug.py --input_image_dir ../images/ --out_dir ../results/

3. Python

NOTE : Dialation and Smudge distort the text on the doc, and is useful for object detection task, Do not use this for OCR or text related task

• Dialation

AugImage = DocumentAugmention.Dialate(BGRImage)

• Smudge

AugImage = DocumentAugmention.Smudge(BGRImage)

• Color Changer
  Color changer will pick random color from 6 distint color and change it with the white background

AugImage = DocumentAugmention.changeColor(image)

• Brightness Up and Down
  
  

Increase the Brightness.

AugImage = DocumentAugmention.BrightnessUp(BGRImage)
AugImage = DocumentAugmention.BrightnessUp(BGRImage,alpha=50.0)

Decrease the Brightness.

AugImage = DocumentAugmention.BrightnessDown(BGRImage)
AugImage = DocumentAugmention.BrightnessDown(BGRImage,alpha=50.0)

Contact

Ayan Gadpal : ayangadpal2 [at] gmail [dot] com

More

Know more about Dialation and smudge in Here