pyiqa 0.1.4

PyTorch Toolbox for Image Quality Assessment

PyTorch Toolbox for Image Quality Assessment

An IQA toolbox with pure python and pytorch. Please refer to Awesome-Image-Quality-Assessment for a comprehensive survey of IQA methods, as well as download links for IQA datasets.

:open_book: Introduction

This is a image quality assessment toolbox with pure python and pytorch. We provide the following features:

• :sparkles: Comprehensive. Support many mainstream full reference (FR) and no reference (NR) metrics
• :sparkles: Accurate. Results calibration of our implementation with official matlab scripts (if exist).
• :sparkles: Fast. With GPU acceleration, most of our implementation is much faster than Matlab.
• :sparkles: Flexible. Support training new DNN models with several public IQA datasets
• :sparkles: Differentiable. Most methods support pytorch backward

Below are details of supported methods and datasets in this project.

Supported methods and datasets:

FR Method Backward PieAPP :white_check_mark: LPIPS :white_check_mark: DISTS :white_check_mark: WaDIQaM :white_check_mark: CKDN1 :white_check_mark: FSIM :white_check_mark: SSIM :white_check_mark: MS-SSIM :white_check_mark: CW-SSIM :white_check_mark: PSNR :white_check_mark: VIF :white_check_mark: GMSD :white_check_mark: NLPD :white_check_mark: VSI :white_check_mark: MAD :white_check_mark:

NR Method Backward FID :heavy_multiplication_x: MANIQA :white_check_mark: MUSIQ :white_check_mark: DBCNN :white_check_mark: PaQ-2-PiQ :white_check_mark: HyperIQA :white_check_mark: NIMA :white_check_mark: WaDIQaM :white_check_mark: CNNIQA :white_check_mark: NRQM(Ma)2 :heavy_multiplication_x: PI(Perceptual Index) :heavy_multiplication_x: BRISQUE :white_check_mark: ILNIQE :white_check_mark: NIQE :white_check_mark:

Dataset Type FLIVE(PaQ-2-PiQ) NR SPAQ NR/mobile AVA NR/Aesthetic PIPAL FR BAPPS FR PieAPP FR KADID-10k FR KonIQ-10k(++) NR LIVEChallenge NR LIVEM FR LIVE FR TID2013 FR TID2008 FR CSIQ FR

[1] This method use distorted image as reference. Please refer to the paper for details.
[2] Currently, only naive random forest regression is implemented and does not support backward.

---

:triangular_flag_on_post: Updates/Changelog

• June 3, 2022. Add FID metric. See clean-fid for more details.
• March 11, 2022. Add pretrained DBCNN, NIMA, and official model of PieAPP, paq2piq.
• More

---

:hourglass_flowing_sand: TODO List

• :white_large_square: Add pretrained models on different datasets.

---

:zap: Quick Start

Dependencies and Installation

• Ubuntu >= 18.04
• Python >= 3.8
• Pytorch >= 1.8.1
• CUDA >= 10.1 (if use GPU)

# Install with pip
pip install pyiqa

# Install latest github version
pip install git+https://github.com/chaofengc/IQA-PyTorch.git

# Install with git clone
git clone https://github.com/chaofengc/IQA-PyTorch.git
cd IQA-PyTorch
pip install -r requirements.txt
python setup.py develop

Basic Usage

import pyiqa
import torch

# list all available metrics
print(pyiqa.list_models())

# create metric with default setting
iqa_metric = pyiqa.create_metric('lpips', device=torch.device('cuda'))
# Note that gradient propagation is disabled by default. set as_loss=True to enable it as a loss function.
iqa_loss = pyiqa.create_metric('lpips', device=torch.device('cuda'), as_loss=True)

# create metric with custom setting
iqa_metric = pyiqa.create_metric('psnr', test_y_channel=True, color_space='ycbcr').to(device)

# check if lower better or higher better
print(iqa_metric.lower_better)

# example for iqa score inference
# Tensor inputs, img_tensor_x/y: (N, 3, H, W), RGB, 0 ~ 1
score_fr = iqa_metric(img_tensor_x, img_tensor_y)
score_nr = iqa_metric(img_tensor_x)

# img path as inputs.
score_fr = iqa_metric('./ResultsCalibra/dist_dir/I03.bmp', './ResultsCalibra/ref_dir/I03.bmp')

# For FID metric, use directory or precomputed statistics as inputs
# refer to clean-fid for more details: https://github.com/GaParmar/clean-fid
fid_metric = pyiqa.create_metric('fid')
score = fid_metric('./ResultsCalibra/dist_dir/', './ResultsCalibra/ref_dir')
score = fid_metric('./ResultsCalibra/dist_dir/', dataset_name="FFHQ", dataset_res=1024, dataset_split="trainval70k")

Example Test Script

Example test script with input directory and reference directory. Single image is also supported for -i and -r options.

# example for FR metric with dirs
python inference_iqa.py -m LPIPS[or lpips] -i ./ResultsCalibra/dist_dir -r ./ResultsCalibra/ref_dir

# example for NR metric with single image
python inference_iqa.py -m brisque -i ./ResultsCalibra/dist_dir/I03.bmp

:hammer_and_wrench: Train

Dataset Preparation

• You only need to unzip downloaded datasets from official website without any extra operation. And then make soft links of these dataset folder under datasets/ folder. Download links are provided in Awesome-Image-Quality-Assessment.
• We provide common interface to load these datasets with the prepared meta information files and train/val/test split files, which can be downloaded from download_link and extract them to datasets/ folder.

You may also use the following commands:

mkdir datasets && cd datasets

# make soft links of your dataset
ln -sf your/dataset/path datasetname

# download meta info files and train split files
wget https://github.com/chaofengc/IQA-PyTorch/releases/download/v0.1-weights/data_info_files.tgz
tar -xvf data_info_files.tgz

Examples to specific dataset options can be found in ./options/default_dataset_opt.yml. Details of the dataloader inferface and meta information files can be found in Dataset Preparation

Example Train Script

Example to train DBCNN on LIVEChallenge dataset

# train for single experiment
python pyiqa/train.py -opt options/train/DBCNN/train_DBCNN.yml

# train N splits for small datasets
python pyiqa/train_nsplits.py -opt options/train/DBCNN/train_DBCNN.yml

:1st_place_medal: Benchmark Performances and Model Zoo

Results Calibration

Please refer to the results calibration to verify the correctness of the python implementations compared with official scripts in matlab or python.

Performances of classical metrics

Here is an example script to get performance benchmark on different datasets:

# NOTE: this script will test ALL specified metrics on ALL specified datasets
# Test default metrics on default datasets
python benchmark_results.py -m psnr ssim -d csiq tid2013 tid2008

# Test with your own options
python benchmark_results.py -m psnr --data_opt options/example_benchmark_data_opts.yml

python benchmark_results.py --metric_opt options/example_benchmark_metric_opts.yml tid2013 tid2008

python benchmark_results.py --metric_opt options/example_benchmark_metric_opts.yml --data_opt options/example_benchmark_data_opts.yml

Please refer to FR benchmark results and NR benchmark results for benchmark performances of some metrics.

Performances of deep learning models

We report PLCC/SRCC here.

Small datasets, validation of split 1

Methods	CSIQ	TID2008	TID2013	LIVE	LIVEM	LIVEC
DBCNN	0.8965/0.9086	0.8322/0.8463	0.7985/0.8320	0.9418/0.9308	0.9461/0.9371	0.8375/0.8530

Large dataset performance

Methods	Dataset	Kon10k	LIVEC	SPAQ	AVA	Link(pth)

:beers: Contribution

Any contributions to this repository are greatly appreciated. Please follow the contribution instructions for contribution guidance.

:scroll: License

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.

:heart: Acknowledgement

The code architecture is borrowed from BasicSR. Several implementations are taken from: IQA-optimization, Image-Quality-Assessment-Toolbox, piq, piqa, clean-fid

We also thanks the following public repositories: MUSIQ, DBCNN, NIMA, HyperIQA, CNNIQA, WaDIQaM, PieAPP, paq2piq, MANIQA

:e-mail: Contact

If you have any questions, please email [e-mail]