Annotation-Augmentation-Tool

Graphical image annotation and augmentation tool for object detection dataset.

Requirments & Installation

This tool is written in python 2.7 based on OpenCV, PyQt4 and numpy libraries. Tested successfully on ubuntu 16.04 and 18.04.

1- clone repositor

 git clone https://github.com/ibrahimsoliman97/Annotation-Augmentation-Tool.git

2- change directory

 cd Annotation-Augmentation-Tool

3- install requirments

  pip install -r requirements.txt ; sudo apt-get install -y python-qt4

4- run the tool

  python AnnotationTool.py

Get Started

	Provides a dialog that allow users to select image or directories contacting our dataset images for starting of annotation process.
	Navigating throw images dataset that located at same chosen directory.
	Save all annotated bounding boxes to a text file in YOLO format.
	List of all available class stated in classes.txt file.
	Perform a rotation augmentation by rotating the image and its annotation by the following degrees (90, 180, 270) and save it in same directory.
	Perform a vertical flipping to the original image.
	Add noise by 2 intensities to the original image with different mean and standard division and save it in same directory.
	Augment original image by generating 2 different brighter images and save it in same directory.

Delete an existing bounding box by right click on it.

Augmentation sample

Orignal Image	Rotated by 90	Rotated by 180	Rotated by 270
Flipped Image	Noise Filter	Brightness level 1	Brightness level 2

Acknowledgement

The author would like to thank the developers of opencv and PyQt4. As well as i would like to acknowledge and express my utmost gratitude to my supervisor Professor Dr. Zulkalnain Bin Mohd Yussof from Faculty of Electronics and Computer Engineering, Universiti Teknikal Malaysia Melaka (UTeM)

The equipment used in this work is provided by Machine Learning and Signal Processing Research Lab, Faculty of Electronic and Computer Engineering, Universiti Technical Malaysia Melaka (UTeM) .