转:http://blog.sina.com.cn/s/blog_4b892b790102vdu3.html#cmt_57105D45-7F000001-E0DF4774-968-8A0
TID2013:
http://www.ponomarenko.info/tid2013.htm
LIVE:
http://live.ece.utexas.edu/research/quality/
GLCM
http://www.fp.ucalgary.ca/mhallbey/tutorial.htm
TID2013旨在用于全参考图像质量的视觉评价。 TID2013主要是评价图像质量评价模型与平均平均的人类感知的匹配。例如,在TID2013中, PSNR(峰值信噪比)和平均人类感知(MOS,平均意见得分)的Spearman相关系数为0.69。
以教育和研究为目的,可以使用,复制或修改本数据库及其文档,无需授权和付费,但要标注出处。
这个数据库在未取得作者的许可的情况下不能修改。作者不保证数据库适用于所有情况。
根据TID2013获得数据结果并发表,请参考 以下论文(见"papers\"目录下的euvip_tid2013.pdf文件):
[1] N. Ponomarenko, O. Ieremeiev, V. Lukin, K. Egiazarian, L. Jin, J. Astola, B. Vozel, K. Chehdi, M. Carli, F. Battisti, C.-C. Jay Kuo, "Color Image Database TID2013: Peculiarities and Preliminary Results", accepted to 4th Europian Workshop on Visual Information Processing EUVIP2013, Paris,
France, June 10-12, 2013, 6p.
TID2013包含25个参考图像和3000的失真图像(25参考图像X24种失真×5失真水平)。所有图像都以Bitmap格式保存在数据库中,没有任何压缩。文件名命名方式,参考图像号,失真类型,失真水平:“iXX_YY_Z.bmp”。
例如,名称为“i03_08_4.bmp”是指第3个参考图像,第8个类型的失真类型与此失真的4种水平。同样地,名称为“i12_10_1.bmp”指这是第12个参考图像,第10个类型的失真与第一种失真水平。“i17.bmp”表示,这是无失真的17个参考图像。
表1 用于TID2013失真的类型
序号 失真类型
1 Additive Gaussian noise
2 Additive noise in color components is more intensive than additive noise in the luminance component
3 Spatially correlated noise
4 Masked noise
5 High frequency noise
6 Impulse noise
7 Quantization noise
8 Gaussian blur
9 Image denoising
10 JPEG compression
11 JPEG2000 compression
12 JPEG transmission errors
13 JPEG2000 transmission errors
14 Non eccentricity pattern noise
15 Local block-wise distortions of different intensity
16 Mean shift (intensity shift)
17 Contrast change
18 Change of color saturation
19 Multiplicative Gaussian noise
20 Comfort noise
21 Lossy compression of noisy images
22 Image color quantization with dither
23 Chromatic aberrations
24 Sparse sampling and reconstruction
详见[1]。
文件“mos.txt”中包含每个图像失真的MOS值。
文件“mos_with_names.txt”包含每个图像失真的MOS值及相应的深圳图像文件名。
文件“mos_std.txt”中包含的每失真图像的MOS的标准差 。
MOS值由971个实验观察者获得的,这观察者来自五个国家:芬兰,法国,意大利,乌克兰和美国。(116实验在芬兰进行,在法国72,80,在意大利,602在乌克兰和美国101)。971个观察者都总共进行了524340次失真图像的对比实验,或者说1048680次图像对的相对质量评价。
MOS(0最小,9最大)的值越大与图像质量越好。
以下文件包含的计算一些质量指标值
在TID2013图像:
“psnrc.txt” - 峰信噪比;
“psnr.txt” – 亮度分量计算出的峰值信噪比;
“ssim.txt” – 结构相似度SSIM值[3];
“mssim.txt” - MSSIM的值[4,2]。
“psnrhvs.txt” - PSNR-HVS度量的值[5];
“psnrhvsm.txt” - PSNR-HVS-M指标值[6];
“psnrha.txt” - PSNRHA值[7];
“psnrhma.txt” - PSNRHMA值[7];
“vifp.txt” - 像素域的VIF值[8,3]。
“nqm.txt” – NQM值[9,3];
“wsnr.txt” -WSNR值[10.3];
“vsnr.txt” - VSNR值[11.3]
“fsim.txt” - FSIM值[12];
“fsimc.txt” - FSIM值[12];
[2] Matthew Gaubatz, "Metrix MUX Visual Quality Assessment Package: MSE, PSNR, SSIM, MSSIM, VSNR, VIF, VIFP, UQI, IFC, NQM, WSNR, SNR", http://foulard.ece.cornell.edu/gaubatz/metrix_mux/
[3] Z. Wang, A. Bovik, H. Sheikh, E. Simoncelli, "Image quality assessment: from error visibility to structural similarity", IEEE Transactions on Image Proc., vol. 13, issue 4, pp. 600-612, April, 2004.
[4] Z. Wang, E. P. Simoncelli and A. C. Bovik, "Multi-scale structural similarity for image quality assessment," Invited Paper, IEEE Asilomar Conference on Signals, Systems and Computers, Nov. 2003.
[5] K. Egiazarian, J. Astola, N. Ponomarenko, V. Lukin, F. Battisti, M. Carli, "New full-reference quality metrics based on HVS", CD-ROM Proceedings of the Second International Workshop on Video Processing and Quality Metrics, Scottsdale, USA, 2006, 4 p.
[6] N. Ponomarenko, F. Silvestri, K. Egiazarian, M. Carli, J. Astola, V. Lukin "On between-coefficient contrast masking of DCT basis functions", CD-ROM Proc. of the Third International Workshop on Video Processing and Quality Metrics. - USA, 2007. - 4 p.
[7] N. Ponomarenko, O. Eremeev, Lukin V., K. Egiazarian, M. Carli, "Modified image visual quality metrics for contrast change and mean shift accounting", Proceedings of CADSM, Polyana-Svalyava, 2011, pp. 305-311.
[8] H.R. Sheikh.and A.C. Bovik, "Image information and visual quality," IEEE Transactions on Image Processing, Vol.15, no.2, 2006, pp. 430-444.
[9] Damera-Venkata N., Kite T., Geisler W., Evans B. and Bovik A. "Image Quality Assessment Based on a Degradation Model", IEEE Trans. on Image Processing, Vol. 9, 2000, pp. 636-650.
[10] T. Mitsa and K. Varkur, "Evaluation of contrast sensitivity functions for the formulation of quality measures incorporated in halftoning algorithms", ICASSP '93-V, pp. 301-304.
[11] D.M. Chandler, S.S. Hemami, "VSNR: A Wavelet-Based Visual Signal-to-Noise Ratio for Natural Images", IEEE Transactions on Image Processing, Vol. 16 (9), pp. 2284-2298, 2007.
[12] L. Zhang, X. Mou, D. Zhang, "FSIM: a feature similarity index for image quality assessment", IEEE Transactions on Image Processing, vol. 20, No 5, 2011, pp. 2378--2386.
程序“spearman.exe”和“kendall.exe”计算了所有图像及下表中特定子集的图像的斯皮尔曼和Kendall的相关系数的一整套的TID2008图像,
表II TID2013默认的子集
No. |
Type of distortion |
Noise |
Actual |
Simple |
Exotic |
New |
Color |
Full |
1 |
Additive Gaussian noise |
+ |
+ |
+ |
- |
- |
- |
+ |
2 |
Noise in color comp. |
+ |
- |
- |
- |
- |
+ |
+ |
3 |
Spatially correl. noise |
+ |
+ |
- |
- |
- |
- |
+ |
4 |
Masked noise |
+ |
+ |
- |
- |
- |
- |
+ |
5 |
High frequency noise |
+ |
+ |
- |
- |
- |
- |
+ |
6 |
Impulse noise |
+ |
+ |
- |
- |
- |
- |
+ |
7 |
Quantization noise |
+ |
- |
- |
- |
- |
+ |
+ |
8 |
Gaussian blur |
+ |
+ |
+ |
- |
- |
- |
+ |
9 |
Image denoising |
+ |
+ |
- |
- |
- |
- |
+ |
10 |
JPEG compression |
- |
+ |
+ |
- |
- |
+ |
+ |
11 |
JPEG2000 compression |
- |
+ |
- |
- |
- |
- |
+ |
12 |
JPEG transm. errors |
- |
- |
- |
+ |
- |
- |
+ |
13 |
JPEG2000 transm. errors |
- |
- |
- |
+ |
- |
- |
+ |
14 |
Non ecc. patt. noise |
- |
- |
- |
+ |
- |
- |
+ |
15 |
Local block-wise dist. |
- |
- |
- |
+ |
- |
- |
+ |
16 |
Mean shift |
- |
- |
- |
+ |
- |
- |
+ |
17 |
Contrast change |
- |
- |
- |
+ |
- |
- |
+ |
18 |
Change of color saturation |
- |
- |
- |
- |
+ |
+ |
+ |
19 |
Multipl. Gauss. noise |
+ |
+ |
- |
- |
+ |
- |
+ |
20 |
Comfort noise |
- |
- |
- |
+ |
+ |
- |
+ |
21 |
Lossy compr. of noisy images |
+ |
+ |
- |
- |
+ |
- |
+ |
22 |
Image color quant. w. dither |
- |
- |
- |
- |
+ |
+ |
+ |
23 |
Chromatic aberrations |
- |
- |
- |
+ |
+ |
+ |
+ |
24 |
Sparse sampl. and reconstr. |
- |
- |
- |
+ |
+ |
- |
+ |
该命令格式是是"spearman " 或者 "kendall ".。
命令行示例:
spearman mos.txt ssim.txt
kendall psnr.txt psnr-hvs.txt
用法示例:
kendall.exe mos.txt FSIMc.txt
Noise : 0.722
Actual : 0.742
Simple : 0.792
Exotic : 0.651
New : 0.611
Color : 0.592
Full : 0.666
表III 各模型与MOS值 的Spearman相关系数排序
等级 |
模型 |
Spearman相关排序 |
1 |
FSIMc |
0.851 |
2 |
PSNR-HA |
0.819 |
3 |
PSNR-HMA |
0.813 |
4 |
FSIM |
0.801 |
5 |
MSSIM |
0.787 |
6 |
PSNRc |
0.687 |
7 |
VSNR |
0.681 |
8 |
PSNR-HVS |
0.654 |
9 |
PSNR |
0.640 |
10 |
SSIM |
0.637 |
11 |
NQM |
0.635 |
12 |
PSNR-HVS-M |
0.625 |
13 |
VIFP |
0.608 |
14 |
WSNR |
0.580 |
表IV各模型与MOS值 的Kendall相关系数排序
等级 |
模型 |
Kendall相关 |
1 |
FSIMc |
0.667 |
2 |
PSNR-HA |
0.643 |
3 |
PSNR-HMA |
0.632 |
4 |
FSIM |
0.630 |
5 |
MSSIM |
0.608 |
6 |
VSNR |
0.508 |
7 |
PSNR-HVS |
0.508 |
8 |
PSNRc |
0.496 |
9 |
PSNR-HVS-M |
0.482 |
10 |
PSNR |
0.470 |
11 |
NQM |
0.466 |
12 |
SSIM |
0.464 |
13 |
VIFP |
0.457 |
14 |
WSNR |
0.446 |
我们计划定期更新这个数据库的版本。新版本
将包括新的类型的失真,并考虑到了另外的实验结果。
如果给我们提供新的模型的及可执行
文件(例如,Matlab代码),将非常感谢(请发至[email protected])。我们保证不会传给其他人员,并将结果纳入到我们的数据库中。