Skip to content
master
Switch branches/tags
Code

Latest commit

 

Git stats

Files

Permalink
Failed to load latest commit information.
Type
Name
Latest commit message
Commit time
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Fast Perceptual Image Enhancement [Paper]

Prerequisites

First steps

  • Download the pre-trained VGG-19 model and put it into vgg_pretrained/ folder
  • Download DPED dataset (patches for CNN training) and extract it into dped/ folder.
    This folder should contain three subolders: sony/, iphone/ and blackberry/, but only iphone/ is needed.

Train the model

python train_model.py model=iphone num_train_iters=40000 run=replication convdeconv depth=16

Test the provided pre-trained model

python test_model.py model=iphone_orig test_subset=full resolution=orig use_gpu=true

Optional parameters:

test_subset: full,small   -   all 29 or only 5 test images will be processed
resolution: orig,high,medium,small,tiny   -   the resolution of the test images [orig means original resolution]
use_gpu: true,false   -   run models on GPU or CPU

Test the obtained model

python test_model.py model=iphone iteration=[40000] test_subset=full resolution=orig use_gpu=true run=replication convdeconv depth=16

Optional parameters:

test_subset: full,small   -   all 29 or only 5 test images will be processed
iteration: all or <number>   -   get visual results for all iterations or for the specific iteration,
               <number> must be a multiple of eval_step
resolution: orig,high,medium,small,tiny   -   the resolution of the test images [orig means original resolution]
use_gpu: true,false   -   run models on GPU or CPU

Folder structure

dped/   -   the folder with the DPED dataset
models/   -   logs and models that are saved during the training process
models_orig/   -   the provided pre-trained models for iphone, sony and blackberry
results/   -   visual results for small image patches that are saved while training
summaries/   -   TensorBoard summaries generated while training
vgg-pretrained/   -   the folder with the pre-trained VGG-19 network
visual_results/   -   processed [enhanced] test images

load_dataset.py   -   python script that loads training data
models.py   -   architecture of the image enhancement [resnet] and adversarial networks
ssim.py   -   implementation of the ssim score
train_model.py   -   implementation of the training procedure
test_model.py   -   applying the pre-trained models to test images
utils.py   -   auxiliary functions
vgg.py   -   loading the pre-trained vgg-19 network

About

Code for "Fast Perceptual Image Enhancement" paper

Resources

Releases

No releases published

Packages

No packages published

Languages