piq 0.5.5

Measures and metrics for image2image tasks. PyTorch.

PyTorch Image Quality (PIQ) is not endorsed by Facebook, Inc.;

PyTorch, the PyTorch logo and any related marks are trademarks of Facebook, Inc.

PyTorch Image Quality (PIQ) is a collection of measures and metrics for image quality assessment. PIQ helps you to concentrate on your experiments without the boilerplate code. The library contains a set of measures and metrics that is continually getting extended. For measures/metrics that can be used as loss functions, corresponding PyTorch modules are implemented.

We provide:

• Unified interface, which is easy to use and extend.

• Written on pure PyTorch with bare minima of additional dependencies.

• Extensive user input validation. You code will not crash in the middle of the training.

• Fast (GPU computations available) and reliable.

• Most metrics can be backpropagated for model optimization.

• Supports python 3.6-3.8.

PIQ was initially named PhotoSynthesis.Metrics.

Installation

PyTorch Image Quality (PIQ) can be installed using pip, conda or git.

If you use pip, you can install it with:

$ pip install piq

If you use conda, you can install it with:

$ conda install piq -c photosynthesis-team -c conda-forge -c PyTorch

If you want to use the latest features straight from the master, clone PIQ repo:

git clone https://github.com/photosynthesis-team/piq.git
cd piq
python setup.py install

Documentation

The full documentation is available at https://piq.readthedocs.io.

Usage Examples

Image-based metrics

The group of metrics (such as PSNR, SSIM, BRISQUE) takes image or images as input. We have a functional interface, which returns a metric value, and a class interface, which allows to use any metric as a loss function.

import torch
from piq import ssim, SSIMLoss

x = torch.rand(4, 3, 256, 256, requires_grad=True)
y = torch.rand(4, 3, 256, 256)

ssim_index: torch.Tensor = ssim(x, y, data_range=1.)

loss = SSIMLoss(data_range=1.)
output: torch.Tensor = loss(x, y)
output.backward()

For a full list of examples, see image metrics examples.

Feature-based metrics

The group of metrics (such as IS, FID, KID) takes a list of image features. Image features can be extracted by some feature extractor network separately or by using the compute_feats method of a class.

Note:

compute_feats consumes a data loader of a predefined format.

import torch
from torch.utils.data import DataLoader
from piq import FID

first_dl, second_dl = DataLoader(), DataLoader()
fid_metric = FID()
first_feats = fid_metric.compute_feats(first_dl)
second_feats = fid_metric.compute_feats(second_dl)
fid: torch.Tensor = fid_metric(first_feats, second_feats)

If you already have image features, use the class interface for score computation:

import torch
from piq import FID

x_feats = torch.rand(10000, 1024)
y_feats = torch.rand(10000, 1024)
msid_metric = MSID()
msid: torch.Tensor = msid_metric(x_feats, y_feats)

For a full list of examples, see feature metrics examples.

List of metrics

Full Reference

Acronym	Year	Metric
PSNR	-	Peak Signal-to-Noise Ratio
SSIM	2003	Structural Similarity
MS-SSIM	2004	Multi-Scale Structural Similarity
VIFp	2004	Visual Information Fidelity
FSIM	2011	Feature Similarity Index Measure
IW-PSNR	2011	Information Weighted PSNR
IW-SSIM	2011	Information Weighted SSIM
SR-SIM	2012	Spectral Residual Based Similarity
GMSD	2013	Gradient Magnitude Similarity Deviation
VSI	2014	Visual Saliency-induced Index
-	2016	Content Score
-	2016	Style Score
HaarPSI	2016	Haar Perceptual Similarity Index
MDSI	2016	Mean Deviation Similarity Index
MS-GMSD	2017	Multi-Scale Gradient Magnitude Similiarity Deviation
LPIPS	2018	Learned Perceptual Image Patch Similarity
PieAPP	2018	Perceptual Image-Error Assessment through Pairwise Preference
DISTS	2020	Deep Image Structure and Texture Similarity

No Reference

Acronym	Year	Metric
TV	1937	Total Variation
BRISQUE	2012	Blind/Referenceless Image Spatial Quality Evaluator

Feature based

Acronym	Year	Metric
IS	2016	Inception Score
FID	2017	Frechet Inception Distance
GS	2018	Geometry Score
KID	2018	Kernel Inception Distance
MSID	2019	Multi-Scale Intrinsic Distance
PR	2019	Improved Precision and Recall

Benchmark

As part of our library we provide code to benchmark all metrics on a set of common Mean Opinon Scores databases. Currently only TID2013 and KADID10k are supported. You need to download them separately and provide path to images as an argument to the script.

Here is an example how to evaluate SSIM and MS-SSIM metrics on TID2013 dataset:

python3 tests/results_benchmark.py --dataset tid2013 --metrics SSIM MS-SSIM --path ~/datasets/tid2013 --batch_size 16

We report Spearman’s Rank Correlation cCoefficient (SRCC) and Kendall rank correlation coefficient (KRCC). We do not report Pearson linear correlation coefficient (PLCC) as it’s highly dependent on fitting method and is biased towards simple examples.

For metrics that can take greyscale or colour images, c means chromatic version.

	TID2013	KADID10k
Acronym	SRCC / KRCC (PIQ)	SRCC / KRCC	SRCC / KRCC (PIQ)	SRCC / KRCC
PSNR	0.6869 / 0.4958	0.687 0.496 TID2013	0.6757 / 0.4876	- / -
SSIM	0.7201 / 0.5271	0.637 / 0.464 TID2013	0.7242 / 0.5370	0.718 / 0.532 KADID10k
MS-SSIM	0.7983 / 0.5965	0.787 / 0.608 TID2013	0.8020 / 0.6088	0.802 / 0.609 KADID10k
VIFp	0.6102 / 0.4579	0.610 / 0.457 TID2013	0.6500 / 0.4770	0.650 / 0.477 KADID10k
FSIM	0.8015 / 0.6289	0.801 / 0.630 TID2013	0.8294 / 0.6390	0.829 / 0.639 KADID10k
FSIMc	0.8509 / 0.6665	0.851 / 0.667 TID2013	0.8537 / 0.6650	0.854 / 0.665 KADID10k
IW-PSNR	- / -	0.6913 / - Eval2019	- / -	- / -
IW-SSIM	- / -	0.7779 / 0.5977 Eval2019	- / -	- / -
SR-SIM	- / -	0.8076 / 0.6406 Eval2019	- / -	0.839 / 0.652 KADID10k
SR-SIMc	- / -	- / -	- / -	- / -
GMSD	0.8038 / 0.6334	0.8030 / 0.6352 MS-GMSD	0.8474 / 0.6640	0.847 / 0.664 KADID10k
VSI	0.8949 / 0.7159	0.8965 / 0.7183 Eval2019	0.8780 / 0.6899	0.861 / 0.678 KADID10k
Content	0.7049 / 0.5173	- / -	0.7237 / 0.5326	- / -
Style	0.5384 / 0.3720	- / -	0.6470 / 0.4646	- / -
HaarPSI	0.8732 / 0.6923	0.8732 / 0.6923 HaarPSI	0.8849 / 0.6995	0.885 / 0.699 KADID10k
MDSI	0.8899 / 0.7123	0.8899 / 0.7123 MDSI	0.8853 / 0.7023	0.885 / 0.702 KADID10k
MS-GMSD	0.8121 / 0.6455	0.8139 / 0.6467 MS-GMSD	0.8523 / 0.6692	- / -
MS-GMSDc	0.8875 / 0.7105	0.687 / 0.496 MS-GMSD	0.8697 / 0.6831	- / -
LPIPS-VGG	0.6696 / 0.4970	0.670 / 0.497 DISTS	0.7201 / 0.5313	- / -
PieAPP	0.8355 / 0.6495	0.875 / 0.710 DISTS	0.8655 / 0.6758	- / -
DISTS	0.8051 / 0.6133	0.830 / 0.639 DISTS	0.8749 / 0.6947	- / -

Assertions

In PIQ we use assertions to raise meaningful messages when some component doesn’t receive an input of the expected type. This makes prototyping and debugging easier, but it might hurt the performance. To disable all checks, use the Python -O flag: python -O your_script.py

Roadmap

See the open issues for a list of proposed features and known issues.

Contributing

If you would like to help develop this library, you’ll find more information in our contribution guide.

Citation

If you use PIQ in your project, please, cite it as follows.

@misc{piq,
  title={{PyTorch Image Quality}: Metrics and Measure for Image Quality Assessment},
  url={https://github.com/photosynthesis-team/piq},
  note={Open-source software available at https://github.com/photosynthesis-team/piq},
  author={Sergey Kastryulin and Dzhamil Zakirov and Denis Prokopenko},
  year={2019},
}

Contacts

Sergey Kastryulin - @snk4tr - [e-mail]

Djamil Zakirov - @zakajd - [e-mail]

Denis Prokopenko - @denproc - [e-mail]