Image Quality

Image quality assessment (IQA) is a multidimensional research problem and an active and evolving research area.
IQA methods can be categorized as subjective versus objective. Subjective methods are based on psycho-physical experiments involving human observers. The human observers assign quality scores from an integer scale. These scores characterize the quality of the images and are related to semantic terms [excellent, good, fair, poor, bad]. In this way, the human observers make a classification, while the given score could be interpreted as a confidence of the observer in her/his judgment.
Objective metrics aim to substitute the subjective experiments. However, it is mandatory to compare the objective scores with the subjective ones. Objective image quality approaches can be categorized into three groups depending on the availability of the original image. Full Reference (FR) methods perform a direct comparison between the image under test and a reference or original image. No Reference (NR) metrics, also called blind methods, are applied when the original image is unavailable. Reduced Reference (RR) metrics lie between FR and NR metrics and are designed to predict image quality with only partial information about the reference image.

IVL DATABASE

The IVL database originates from 20 reference images of 886×591 pixels (15×10 cm at 150 dpi, typical printing parameters for natural photos), chosen to sample different contents both in terms of low level features (frequencies, colors) and higher ones (face, buildings, close-up, outdoor, landscape).
Starting from these images we have generated:
1] A database of single distortions SD-IVL: 200 noisy images and 180 JPEG compressed images.
2] A database of multiply distorted images MD-IVL: 400 images with noise plus JPEG distrotions and 350 images with blur plus JPEG distortions.
If you use the single distortion database SD-IVL in your research, we kindly ask you to cite this paper:
@article{corchs2014no-reference,
 author = {Corchs, Silvia and Gasparini, Francesca and Schettini, Raimondo},
 year = {2014},
 pages = {86-100},
 title = {No Reference Image Quality classification for JPEG-Distorted Images},
 volume = {30},
 publisher = {Elsevier},
 journal = {Digital Signal Processing},
 doi = {10.1016/j.dsp.2014.04.003},
 issn = {1051-2004}
}
Database is password protected, please contact gasparini-at-disco.unimib.it and the password will be sent to you.
The use of these datasets is intended only for accademic or educational purposes.

 

DATASETS

Single Distortion Database SD-IVL

  Subjective Data IVL_SD (17.3 KiB, 81 hits)

  SD_IVL dataset (244.0 MiB, 116 hits)

  copyright notice IVL (239.9 KiB, 230 hits)

 

Multiply Distorted Database MD -IVL

  MDIVL.zip (95.3 MiB, 78 hits)

  copyright notice MDIVL (399.6 KiB, 153 hits)

 

Publications

1.

Quality of Images
(Gianluigi Ciocca, Silvia Corchs, Francesca Gasparini, Carlo Batini, Raimondo Schettini) Chapter in Data and Information Quality: Dimensions, Principles and Techniques, Springer, 2015.

@inbook{ciocca2015quality-image,
 author = {Ciocca, Gianluigi and Corchs, Silvia and Gasparini, Francesca and Batini, Carlo and Schettini, Raimondo},
 editor = {Batini, Carlo and Scannapieco, Monica},
 year = {2015},
 title = {Quality of Images},
 publisher = {Springer},
 booktitle = {Data and Information Quality: Dimensions, Principles and Techniques},
 doi = {10.1007/978-3-319-24106-7}}
2.

Modeling Image Quality
(Gianluigi Ciocca, Silvia Corchs, Francesca Gasparini, Raimondo Schettini) Chapter in Encyclopedia of Information Science and Technology, pp. 569-580, IGI Global, 2014.

@inbook{ciocca2013modeling-image,
 author = {Ciocca, Gianluigi and Corchs, Silvia and Gasparini, Francesca and Schettini, Raimondo},
 editor = {Khosrow-Pour, Mehdi},
 year = {2014},
 pages = {569-580},
 title = {Modeling Image Quality},
 publisher = {IGI Global},
 booktitle = {Encyclopedia of Information Science and Technology},
 doi = {10.4018/978-1-4666-5888-2.ch590}}
3.

No Reference Image Quality classification for JPEG-Distorted Images
(Silvia Corchs, Francesca Gasparini, Raimondo Schettini) In Digital Signal Processing, volume 30, pp. 86-100, Elsevier, 2014.

@article{corchs2014no-reference,
 author = {Corchs, Silvia and Gasparini, Francesca and Schettini, Raimondo},
 year = {2014},
 pages = {86-100},
 title = {No Reference Image Quality classification for JPEG-Distorted Images},
 volume = {30},
 publisher = {Elsevier},
 journal = {Digital Signal Processing},
 url = {http://www.sciencedirect.com/science/article/pii/S1051200414001080},
 doi = {10.1016/j.dsp.2014.04.003},
 issn = {1051-2004},
 projectref = {http://www.ivl.disco.unimib.it/research/ImageQuality/}}
4.

Noisy Images-JPEG Compressed: Subjective and Objective Image Quality Evaluation
(Silvia Corchs, Francesca Gasparini, Raimondo Schettini) In Image Quality and System Performance XI, volume 9016, pp. 90160-, SPIE, 2014.

@inproceedings{corchs2014noisy-images,
 author = {Corchs, Silvia and Gasparini, Francesca and Schettini, Raimondo},
 year = {2014},
 pages = {90160-},
 title = {Noisy Images-JPEG Compressed: Subjective and Objective Image Quality Evaluation},
 volume = {9016},
 publisher = {SPIE},
 booktitle = {Image Quality and System Performance XI},
 pdf = {/download/corchs2014noisy-images.pdf},
 doi = {10.1117/12.2039273},
 projectref = {http://www.ivl.disco.unimib.it/research/ImageQuality/}}
5.

How to assess image quality within a workflow chain: an overview
(Gianluigi Ciocca, Silvia Corchs, Francesca Gasparini, Raimondo Schettini) In International Journal on Digital Libraries, volume 15, number 1, pp. 1-25, Springer, 2014.

@article{ciocca2014how-assess,
 author = {Ciocca, Gianluigi and Corchs, Silvia and Gasparini, Francesca and Schettini, Raimondo},
 year = {2014},
 pages = {1-25},
 title = {How to assess image quality within a workflow chain: an overview},
 volume = {15},
 number = {1},
 publisher = {Springer},
 journal = {International Journal on Digital Libraries},
 doi = {10.1007/s00799-014-0124-0}}
6.

Grouping strategies to improve the correlation between subjective and objective image quality data
(Silvia Corchs, Francesca Gasparini, Raimondo Schettini) In Image Quality and System Performance X, IS&T/SPIE Electronic Imaging, volume 8653, pp. 86530d(1-8), SPIE, 2013.

@inproceedings{corchs2013grouping-strategies,
 author = {Corchs, Silvia and Gasparini, Francesca and Schettini, Raimondo},
 year = {2013},
 pages = {86530d(1-8)},
 title = {Grouping strategies to improve the correlation between subjective and objective image quality data},
 volume = {8653},
 publisher = {SPIE},
 booktitle = {Image Quality and System Performance X, IS\&T/SPIE Electronic Imaging},
 pdf = {/download/corchs2013grouping-strategies.pdf},
 doi = {10.1117/12.2006227},
 projectref = {http://www.ivl.disco.unimib.it/research/ImageQuality/}}
7.

A no-reference metric for demosaicing artifacts that fits psycho-visual experiments
(Francesca Gasparini, Fabrizio Marini, Raimondo Schettini, Mirko Guarnera) In EURASIP Journal on Advances in Signal Processing, volume 2012, number 1, pp. 4868-4873, 2012.

@article{gasparini2012No-reference-metric,
 author = {Gasparini, Francesca and Marini, Fabrizio and Schettini, Raimondo and Guarnera, Mirko},
 year = {2012},
 pages = {4868-4873},
 title = {A no-reference metric for demosaicing artifacts that fits psycho-visual experiments},
 volume = {2012},
 number = {1},
 journal = {EURASIP Journal on Advances in Signal Processing},
 pdf = {/download/gasparini2012No-reference-metric.pdf}}
8.

Bio-inspired framework for automatic image quality enhancement
(Andrea Ceresi, Francesca Gasparini, Fabrizio Marini, Raimondo Schettini) In Digital Photography VIII, IS&T/SPIE Electronic Imaging, volume 8299, pp. 82990g (1-8), SPIE, 2012.

@inproceedings{ceresi2012bio-inspired-framework,
 author = {Ceresi, Andrea and Gasparini, Francesca and Marini, Fabrizio and Schettini, Raimondo},
 year = {2012},
 pages = {82990g (1-8)},
 title = {Bio-inspired framework for automatic image quality enhancement},
 volume = {8299},
 publisher = {SPIE},
 booktitle = {Digital Photography VIII, IS\&T/SPIE Electronic Imaging},
 pdf = {/download/ceresi2012bio-inspired-framework.pdf},
 doi = {10.1117/12.910712}}
9.

A sharpness measure on automatically selected edge segments
(Silvia Corchs, Francesca Gasparini, Fabrizio Marini, Raimondo Schettini) In Image Quality and System Performance IX, IS&T/SPIE Electronic Imaging, volume 8293, pp. 82930a(1-8), SPIE, 2012.

@inproceedings{corchs2012sharpness-measure,
 author = {Corchs, Silvia and Gasparini, Francesca and Marini, Fabrizio and Schettini, Raimondo},
 year = {2012},
 pages = {82930a(1-8)},
 title = {A sharpness measure on automatically selected edge segments},
 volume = {8293},
 publisher = {SPIE},
 booktitle = {Image Quality and System Performance IX, IS\&T/SPIE Electronic Imaging},
 pdf = {/download/corchs2012sharpness-measure.pdf},
 doi = {10.1117/12.908473}}
10.

Image quality: a tool for no-reference assessment methods
(Silvia Corchs, Francesca Gasparini, Fabrizio Marini, Raimondo Schettini) In Image Quality and System Performance VIII, IS&T/SPIE Electronic Imaging, volume 7867, pp. 78760x (1-9), SPIE, 2011.

@inproceedings{corchs2011image-quality,
 author = {Corchs, Silvia and Gasparini, Francesca and Marini, Fabrizio and Schettini, Raimondo},
 year = {2011},
 pages = {78760x (1-9)},
 title = {Image quality: a tool for no-reference assessment methods},
 volume = {7867},
 publisher = {SPIE},
 booktitle = {Image Quality and System Performance VIII, IS\&T/SPIE Electronic Imaging},
 doi = {10.1117/12.872367}, % project-url = {http://www.ivl.disco.unimib.it/research/ImageQuality/}}
11.

No-reference metrics for demosaicing
(Francesca Gasparini, Mirko Guarnera, Fabrizio Marini, Raimondo Schettini) In Image Quality and System Performance VII, IS&T/SPIE Electronic Imaging, volume 7529, pp. 752911-752911, SPIE, 2010.

@inproceedings{gasparini2010no-reference-metrics,
 author = {Gasparini, Francesca and Guarnera, Mirko and Marini, Fabrizio and Schettini, Raimondo},
 editor = {Susan P. Farnand, Frans Gaykema},
 year = {2010},
 pages = {752911-752911},
 title = {No-reference metrics for demosaicing},
 volume = {7529},
 publisher = {SPIE},
 booktitle = {Image Quality and System Performance VII, IS\&T/SPIE Electronic Imaging},
 doi = {10.1117/12.839952}, %project-url = {http://www.ivl.disco.unimib.it/research/ImageQuality/}}