Download python 3.6
Download pyqt4
Install python
then,
use following commands to install the libraries
$ pip install tensorflow
$ pip install keras
$ pip install scikit-image
Getting Dataset and trained model file
The Train and Test data files are available on the following link.
https://drive.google.com/drive/folders/1ttSWCJGWZdGLeTHRzHHVxuAhJUF7c5An?usp=sharing
Running the tests
Once setup is ready,
Execute Prediction.py
Double click on ColorIt.py
- Click on the "HOME" button
- Click on the "Select a file from computer" button
- Once file is uploaded, click on "Colorise" button
- Result will be saved in a folder on desktop
Train
This model is trained on Google Colab GPU.
This software was trained on 1000 image dataset and we got pretty good results but we can further train it on more images to improvise the results.
Built With
- Python 3
- pyQt4
- Convolustional neural network using keras
Reference
Research paper -
https://arxiv.org/pdf/1712.03400.pdf
http://cs231n.stanford.edu/reports/2016/pdfs/219_Report.pdf
Version
ColorIt version 1.0
Developers
Tanish Sehgal
Jaineel Shah
Pravar Parekh
Atharva Gole
Shivashish Jaiswal
About
Image Colorisation Application which colorises black and white images to their approximate colour images