[go: up one dir, main page]

Skip to content

Project for Deep Learning. Using Generative Adversarial Network (GAN) to colorize grayscale images. Trained on COCO Dataset.

Notifications You must be signed in to change notification settings

FaizaanAlFaisal/Image-Colorization-with-GANs

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 
 
 

Repository files navigation

Image Colorization using GANs

This project utilizes Generative Adversarial Networks to colorize grayscale images.

The training and GPU resources were utilized primarily from Kaggle Notebooks and Google Colab environments.

Dataset

The COCO Dataset is the one used for training and testing purposes.

Model Description

The final model selected utilized a UNet as the Generator, and a Patch Discriminator model as the Discriminator.

Sample Output

The following is a sample output of the model after training. The top row shows input grayscale images, the middle row shows images colorized using the model, and the bottom row show the real colored version of the image.

Sample output of GAN. Top row shows Grayscale image. Middle row shows colored images generated by GAN. Bottom row show the real colored versions of those images.

About

Project for Deep Learning. Using Generative Adversarial Network (GAN) to colorize grayscale images. Trained on COCO Dataset.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published