purvaten / blind-dehazing

An implementation of the ICCP '16 paper "Blind Dehazing Using Internal Patch Recurrence".

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Blind Dehazing Algorithms

This is an implementation of single image blind dehazing algorithms. We attempt to implement two papers, the ICCP '16 paper "Blind dehazing using internal patch recurrence", by Yuval Bahat & Michal Irani, and the CVPR '09 paper "Single Image Haze Removal Using Dark Channel Prior" by Kaiming He, Jian Sun, and Xiaoou Tang.

The implementation of "Single Image Haze Removal Using Dark Channel Prior" is complete. Run it using ./start_dark_channel.sh. Use the --help flag for a detailed list of arguments. Hyperparameters can be adjusted in dark_prior/config/constants.yml.

The implementation of "Blind dehazing using internal patch recurrence" is in progress.

Blind dehazing using internal patch recurrence Results

Transmission Maps




Results before tmap estimation (assuming lower bound on tmap)




Single Image Haze Removal Using Dark Channel Prior Results

You can find some results in this report.

About

An implementation of the ICCP '16 paper "Blind Dehazing Using Internal Patch Recurrence".


Languages

Language:Python 99.5%Language:Shell 0.5%