Skip to content

This project implements a neural network for image colourisation using the popular Unet architecture. While Unet is typically used for image segmentation tasks, it can also be used for image reconstruction and colourisation..

Notifications You must be signed in to change notification settings

Ali-Fayzi/colourisation

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

10 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Unet-based Conditional Image Colourisation

This project implements a neural network for image colourisation using the popular Unet architecture. While Unet is typically used for image segmentation tasks, it can also be used for image reconstruction and colourisation.

Installation

To install the required packages, run:

pip install -r requirements.txt

Usage

To use the trained network to generate colorized images, run:

python inference.py --weight_path ./weights/weight.pt --input_path ./images --output_path ./output --plot

You can specify the path to the trained weights, the input directory containing grayscale images, the output directory to save colorized images, and whether or not to display the images.

Output

     Gray Scale Image		   Random Pixel		       Output	              Ground Truth	

Out1 Out2 Out3 Out4 Out5

Training

The network was trained on the COCO dataset using a customized dataloader that randomly selects only 0.1% of the actual image pixels for conditional colorization. Due to limited resources, the network was trained for only 2 epochs.

Acknowledgements

This project was implemented by Ali Fayzi. Feel free to use and modify the code for your own purposes. If you have any questions or suggestions, please feel free to contact me.

About

This project implements a neural network for image colourisation using the popular Unet architecture. While Unet is typically used for image segmentation tasks, it can also be used for image reconstruction and colourisation..

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages