[STATUS - ACTIVE]

[DIDH] - Towards Domain Invariant Single Image Dehazing

Official Pytorch based implementation.

To-Do List

• Paper Link
• Training Code
• Trained Models
• Trained Models corresponding to experiments
• Citation

Results

Standard Performance Evaluation of different Dehazing algorithms on NTIRE-19 dataset

Blind Evaluation of different Dehazing algorithms on NTIRE-18 dataset

Dependencies and Installation

• python=3.8
• PyTorch=1.6
• tqdm
• numpy
• matplotlib
• PIL
• piq
• OpenCV=4.4

Dataset Details

Dataset Name	PSNR / SSIM	Resolution	Type	Paper Link
I-Haze NTIRE-18	14.60 / 0.67	4177 x 3134	Real	paper
O-Haze NTIRE-18	14.60 / 0.67	4177 x 3134	Real	paper
NTIRE-19	9.11 / 0.49	1600 x 1200	Real	paper
NTIRE-20	10.42 / 0.46	1600 x 1200	Real	paper
SOTS-IN	11.97 / 0.69	620 x 460	Synthetic	paper
SOTS-OUT	15.92 / 0.81	550 x 478	Synthetic	paper
HazeRD	14.60 / 0.67	3492 x 2558	Synthetic	paper

MODEL ZOO - Checkpoints of Networks

Trained on Synthetic (RESIDE-Indoor) dataset -> Trained_Models

Model	SOTS-IN (PSNR / SSIM)	SOTS-OUT (PSNR / SSIM)	NTIRE-19 (PSNR / SSIM)	NTIRE-20 (PSNR / SSIM)
DuRN-US	32.12 / 0.98	19.55 / 0.83	10.81 / 0.51	11.27 / 0.51
FFA-Net	36.36 / 0.98	20.05 / 0.84	10.97 / 0.42	10.70 / 0.44
Wavelet-UNet	20.02 / 0.75	17.75 / 0.67	11.48 / 0.47	10.88 / 0.36
MSNet	32.04 / 0.98	20.70 / 0.86	9.90 / 0.51	11.16 / 0.51
SNDN	24.68 / 0.91	16.02 / 0.69	10.13 / 0.45	10.64 / 0.43
GridDehazeNet	32.14 / 0.98	16.22 / 0.76	9.50 / 0.49	9.01 / 0.40
PFFNet	26.58 / 0.92	14.63 / 0.65	11.38 / 0.51	11.14 / 0.43
Ours	38.91 / 0.98	25.75 / 0.84	16.21 / 0.78	16.28 / 0.67

Trained on Synthetic (RESIDE-Outdoor) dataset -> Trained_Models

Model	SOTS-IN (PSNR / SSIM)	SOTS-OUT (PSNR / SSIM)	NTIRE-19 (PSNR / SSIM)	NTIRE-20 (PSNR / SSIM)
DuRN-US	15.95 / 0.76	19.41 / 0.81	11.04 / 0.51	11.73 / 0.46
FFA-Net	18.96 / 0.86	30.88 / 0.93	9.64 / 0.50	10.90 / 0.48
Wavelet-UNet	16.26 / 0.73	21.95 / 0.76	10.36 / 0.49	11.05 / 0.43
MSNet	21.75 / 0.88	29.80 / 0.93	09.56 / 0.48	11.35 / 0.51
SNDN	25.30 / 0.91	24.31 / 0.88	11.74 / 0.49	11.95 / 0.52
GridDehazeNet	20.99 / 0.89	29.18 / 0.93	10.16 / 0.50	11.23 / 0.49
PFFNet	20.32 / 0.85	27.65 / 0.91	10.75 / 0.50	11.55 / 0.52
Ours	26.90 / 0.76	30.40 / 0.94	13.36 / 0.52	12.68 / 0.52

Trained on Real (NTIRE-19) dataset -> Trained_Models

Model	SOTS-IN (PSNR / SSIM)	SOTS-OUT (PSNR / SSIM)	NTIRE-19 (PSNR / SSIM)	NTIRE-20 (PSNR / SSIM)
DuRN-US	11.44 / 0.59	13.05 / 0.61	13.63 / 0.57	12.97 / 0.52
FFA-Net	12.16 / 0.55	14.36 / 0.59	14.01 / 0.56	14.71 / 0.57
Wavelet-UNet	13.57 / 0.41	13.05 / 0.44	12.85 / 0.39	12.08 / 0.24
MSNet	13.33 / 0.55	13.85 / 0.56	13.32 / 0.53	12.63 / 0.32
SNDN	12.56 / 0.66	14.11 / 0.70	13.54 / 0.54	14.93 / 0.51
GridDehazeNet	14.57 / 0.59	13.47 / 0.60	12.96 / 0.50	12.07 / 0.32
PFFNet	13.51 / 0.50	14.57 / 0.53	13.29 / 0.52	12.99 / 0.31
Ours	19.28 / 0.66	18.17 / 0.87	19.47 / 0.75	20.33 / 0.77

Trained on Real (NTIRE-20) dataset -> Trained_Models

Model	SOTS-IN (PSNR / SSIM)	SOTS-OUT (PSNR / SSIM)	NTIRE-19 (PSNR / SSIM)	NTIRE-20 (PSNR / SSIM)
DuRN-US	9.43 / 0.63	11.92 / 0.66	11.63 / 0.52	15.27 / 0.50
FFA-Net	9.96 / 0.63	14.88 / 0.75	12.43 / 0.52	18.11 / 0.66
Wavelet-UNet	12.04 / 0.32	13.85 / 0.41	11.46 / 0.28	12.08 / 0.21
MSNet	9.16 / 0.51	10.66 / 0.56	12.04 / 0.50	14.06 / 0.50
SNDN	12.03 / 0.67	14.14 / 0.73	11.73 / 0.52	13.93 / 0.52
GridDehazeNet	11.60 / 0.58	12.75 / 0.72	13.39 / 0.52	15.32 / 0.60
PFFNet	8.82 / 0.47	12.00 / 0.53	11.54 / 0.49	14.50 / 0.36
Ours	19.53 / 0.71	18.69 / 0.79	17.24 / 0.66	21.17 / 0.78

Trained on Aggregated dataset -> Trained_Models

Model	SOTS-IN (PSNR / SSIM)	SOTS-OUT (PSNR / SSIM)	NTIRE-19 (PSNR / SSIM)	NTIRE-20 (PSNR / SSIM)
DuRN-US	26.80 / 0.95	29.59 / 0.91	15.96 / 0.61	19.88 / 0.69
FFA-Net	19.15 / 0.85	25.53 / 0.89	14.06 / 0.54	15.93 / 0.59
Wavelet-UNet	15.46 / 0.65	19.21 / 0.67	12.23 / 0.51	12.66 / 0.42
MSNet	24.38 / 0.90	26.79 / 0.90	14.65 / 0.59	15.17 / 0.63
SNDN	22.94 / 0.88	25.49 / 0.89	14.66 / 0.59	18.65 / 0.67
GridDehazeNet	22.87 / 0.91	27.18 / 0.91	13.97 / 0.56	17.02 / 0.68
PFFNet	23.39 / 0.87	25.80 / 0.87	13.50 / 0.48	14.77 / 0.57
Ours	25.39 / 0.80	19.47 / 0.75	22.80 / 0.82	23.66 / 0.88

Blind Evaluation on Haze-RD and NTIRE-18 dataset

Algorithm	Haze-RD (PSNR/SSIM)	NTIRE-18 (PSNR/SSIM)
DuRN-US	15.26 / 0.83	18.85 / 0.71
FFA-Net	15.77 / 0.83	16.16 / 0.64
Wavelet	16.30 / 0.82	16.01 / 0.70
MSNet	14.42 / 0.80	17.42 / 0.66
SNDN	16.05 / 0.82	18.08 / 0.70
GridDehazeNet	14.58 / 0.81	18.26 / 0.74
PFFNet	15.20 / 0.77	15.66 / 0.58
DA-Dehaze	16.21 / 0.78	16.28 / 0.67
Ours	21.42 / 0.81	24.14 / 0.80

Code Details

Create a 1-1 list corresponding to ground truth and hazy image

1. Update location of datasets in gen_paired_datasets.py
2. Comment out datasets that are not needed and run,

python gen_paired_dataset.py

Working of Greedy Localized Data Augmentation

To see how localized data augmentation works,

python dataloader.py

Train

python train.py

Test

python eval.py