Skip to content
This repository has been archived by the owner on Nov 12, 2018. It is now read-only.
/ photo-auto-balance Public archive

🌅 Automated image color correction / enhancement with Deep Learning (CNN) ***(NO LONGER MAINTAINED)

Notifications You must be signed in to change notification settings

tao-pr/photo-auto-balance

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Photo Auto Balancer

A quick fun project which utilises Deep Learning to automate image color/level adjustment.


CHART


Quick Start

Install all prerequisites with the following command.

$ pip3 install -r requirements.txt

Prepare training directory

Say your preferred trainset directory is /path/to/trainset/. Create three subdirectories inside it as follows.

trainset
├── out/
├── raw/
└── unfiltered/

Then put your images as a trainset inside /path/to/trainset/raw. Leave the other two empty. You're all set.

Training

$ python3 loader.py --train --limit 30 --dir /path/to/trainset/

The script reads all JPG images from the dir you specified in the arguments. The reverse filtered images will be generated inside out subdirectory.

CAVEAT: The process starts training Convolutional Neural Network rightaway after the reverse filtered samples are generated. This takes huge computational power and time.

Pro Tip: Quick Setup on Ubuntu

To set all dependencies up and start the training process in one go within just few minutes, run the following script:

$ ./setup-ec2-and-run

What the script does are:

  • Install all required packages
  • Download the primary training data
  • Start the training process in background
  • Leaves the training log at /home/ubuntu/photo-auto-balance/log.txt

Licence

MIT licence.

About

🌅 Automated image color correction / enhancement with Deep Learning (CNN) ***(NO LONGER MAINTAINED)

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published