The DOI matches its Crossref record; 16 of the 16 references it cites were found in a registry or an index; none of the 14 sources its claims cite is checked for support yet.
Multiscale structural similarity for image quality assessment
Abstract
The structural similarity image quality paradigm is based on the assumption that the human visual system is highly adapted for extracting structural information from the scene, and therefore a measure of structural similarity can provide a good approximation to perceived image quality. This paper proposes a multiscale structural similarity method, which supplies more flexibility than previous single-scale methods in incorporating the variations of viewing conditions. We develop an image synthesis method to calibrate the parameters that define the relative importance of different scales. Experimental comparisons demonstrate the effectiveness of the proposed method.
Overview
CoChat Summary
TL;DR
This paper proposes a multi-scale structural similarity (MS-SSIM) image quality method that combines contrast and structure comparisons across scales and luminance comparison at the coarsest scale. On the LIVE JPEG/JPEG2000 database, MS-SSIM outperformed PSNR, the Sarnoff model, and single-scale SSIM variants on all reported evaluation criteria.
Key findings
- MS-SSIM achieved the best reported performance: CC 0.969, ROCC 0.966, MAE 3.86, RMS 4.91, and OR 1.16%, compared with the best single-scale SSIM at M=2 (CC 0.963, ROCC 0.959, MAE 4.21, RMS 5.38, OR 2.62%) and PSNR (CC 0.905, ROCC 0.901, MAE 6.53, RMS 8.45, OR 15.7%).
- Single-scale SSIM performance varied with scale; the best single-scale result was at M=2, and single-scale SSIM tended to give higher scores as scale increased.
- The calibrated cross-scale exponents were β1=γ1=0.0448, β2=γ2=0.2856, β3=γ3=0.3001, β4=γ4=0.2363, and α5=β5=γ5=0.1333.
- The authors conclude that multi-scale SSIM provides a meaningful balance between scales and outperforms the best single-scale SSIM and state-of-the-art image quality metrics.
Methods
- The MS-SSIM system iteratively applies a low-pass filter and downsamples by 2; at each scale it computes contrast and structure comparisons, computes luminance comparison only at Scale M, and combines them using exponents normalized so that ∑γj=1.
- Cross-scale parameters were calibrated by image synthesis: 5 scales and 12 distortion levels (range 2^3 to 2^14) were used with 10 original 64×64 images, producing 600 distorted images; 8 subjects selected images of equal perceived quality across scales at a viewing distance of 32 pixels per degree of visual angle.
- Evaluation used the LIVE database with 344 JPEG and JPEG2000 compressed images (bit rate 0.028 to 3.150 bits/pixel), MOS from 13–25 subjective scores, and compared PSNR, Sarnoff JNDmetrix 8.0, single-scale SSIM for M=1–5, and MS-SSIM using CC, MAE, RMS, OR, and ROCC after logistic mapping.
Limitations
- Stated: the image-synthesis calibration approach is “still rather crude,” and the authors are working on a more systematic approach.
- Stated: CSF cannot be directly used to derive the parameters because it is typically measured at visibility threshold with simplified stimuli, whereas the goal is comparing complex structured images at visible distortion levels.
- Evident from the text: calibration used only 10 original 64×64 grayscale images and 8 subjects, including one author, and evaluation was limited to the LIVE JPEG/JPEG2000 database.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
0 Valid · 1 Not valid · 14 Not checked · 18 Own results and interpretations
A result the paper states — it cites no source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“It can also be observed that the single-scale model tends to supply higher scores with the increase of scales.”
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Finally, for every one of the objective evaluation criteria, multi-scale SSIM model outperforms all the other models, including the best single-scale SSIM model, suggesting a meaningful balance between scales.”
- Also stated in: § 5. DISCUSSIONS · p. 4 · hedged
“Experiments show that with an appropriate parameter settings, the multi-scale method outperforms the best single-scale SSIM model as well as state-of-the-art image quality metrics.”
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Also stated in: § 5. DISCUSSIONS · p. 4 · hedged
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“In general, the subjects agreed with each other on each image more than they agreed with themselves across different images.”
plainstates an associationn = 8- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The resulting parameters we obtained are β 1 = γ 1 = 0.0448, β 2 = γ 2 = 0.2856, β 3 = γ 3 = 0.3001, β 4 = γ 4 = 0.2363, and α 5 = β 5 = γ 5 = 0.1333, respectively.”
plainn = 8- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“However, the SSIM index algorithm introduced in [5] is a single-scale approach.”
What the sources say
- Not matchedcites [5] [5] Z. Wang, A. C. Bovik, H. R. Sheikh, and E. P. Simoncelli, 'Image quality assessment: From error measurement to structural similarity,' IEEE Trans. Image Processing , vol. 13, Jan. 2004. -
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Not matched
10 more claims
Create a free account to see all 33 claims and every source verdict.
- It may state two separate claims as one. 8
- Probably not a substantial claim. 13
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The cited source says what the claim attributes to it.
In the citing paper
“Further, as research in quality assessment has demonstrated, incorporating multiscale information when assessing quality produces QA algorithms that perform better in terms of correlation with human perception [13], [29].”
From No-Reference Image Quality Assessment in the Spatial Domain (2012) · cites this paper as [13]
Citations
Cites
Image quality assessment: from error visibility to structural similarity
Wang, Z., Bovik, A. C., Sheikh, H. R., & Simoncelli, E. P. (2004) · IEEE Transactions on Image Processing · doi:10.1109/tip.2003.819861 · Cited by 57,892
Self-citationA universal image quality index
Wang, Z., & Bovik, A. C. (2002) · IEEE Signal Processing Letters · doi:10.1109/97.995823 · Cited by 5,757
Self-citationImage quality measures and their performance
Eskicioglu, A. M., & Fisher, P. S. (1995) · IEEE Transactions on Communications · doi:10.1109/26.477498 · Cited by 2,117
A Parametric Texture Model Based on Joint Statistics of Complex Wavelet Coefficients
Portilla, J., & Simoncelli, E. P. (2000) · International Journal of Computer Vision · doi:10.1023/a:1026553619983 · Cited by 1,806
Self-citation(1995) · Medical Entomology and Zoology · Cited by 1,212
Pyramid-based texture analysis/synthesis
Heeger, D. J., & Bergen, J. R. (1995) · Proceedings of the 22nd annual conference on Computer graphics and interactive techniques - SIGGRAPH '95 · doi:10.1145/218380.218446 · Cited by 1,095
Video quality assessment based on structural distortion measurement
Wang, Z., Lu, L., & Bovik, A. C. (2004) · Signal Processing: Image Communication · doi:10.1016/s0923-5965(03)00076-6 · Cited by 1,024
Self-citationImage Quality Assessment: From Error Measurement to Structural Similarity
Wang, Z., Bovik, A. C., Sheikh, H. R., & Simoncelli, E. P. (2004) · Cited by 881
Self-citationTeo, P. C., & Heeger, D. J. (2002) · Proceedings of 1st International Conference on Image Processing · doi:10.1109/icip.1994.413502 · Cited by 471
Objective video quality assessment
(2003) · Handbook of Video Databases · doi:10.1201/9780203489864-45 · Cited by 359
Research works printed but not linked: 2
- 0, JNDmetrix Technology, Sarnoff Corporation
- 2000, Final report from the video quality experts group on the validation of objective models of video quality assessment, VQEG
Cited by
Items 2,301–2,400 of 6,047
Guo, Q. J., Qian, Y., Wang, H., Yu, W., Xu, F., Cui, T. J., & Jin, Y. Q. (2022) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2022.3154932 · Cited by 13
Objective Quality Assessment Metrics for Light Field Image Based on Textural Features
Cong, H. P., Perry, S. W., Cheng, E., et al. (2022) · Electronics · doi:10.3390/electronics11050759 · Cited by 11
Usui, K., Ogawa, K., Goto, M., et al. (2022) · Radiation Oncology · doi:10.1186/s13014-022-02042-1 · Cited by 11
Remodelling Correlation: A Fault Resilient Technique of Correlation Sensitive Stochastic Designs
Mitra, S., Banerjee, S., & Naskar, M. K. (2022) · Array · doi:10.1016/j.array.2022.100219 · Cited by 11
A Joint Spatiotemporal Video Compression Based on Stochastic Adaptive Fourier Decomposition
Dai, L., & Zhang, L. (2022) · IEEE Signal Processing Letters · doi:10.1109/lsp.2022.3187917 · Cited by 11
Yu, Y. F., & Qin, C. (2023) · Fractal and Fractional · doi:10.3390/fractalfract7010070 · Cited by 11
Dense-FG: A Fusion GAN Model by Using Densely Connected Blocks to Fuse Infrared and Visible Images
Xu, X., Shen, Y., & Han, S. (2023) · Applied Sciences · doi:10.3390/app13084684 · Cited by 11
Islam, N., & Shin, S. (2023) · Electronics · doi:10.3390/electronics12112425 · Cited by 11
IRGAN: cGAN-based Indoor Radio Map Prediction
Cissé, C. T., Baala, O., Guillet, V., et al. (2023) · IFIP Networking Conference (IFIP Networking) · doi:10.23919/ifipnetworking57963.2023.10186359 · Cited by 11
“Finally, the total loss of our cGAN models is a weighted sum of the above loss functions and is given by the following equation: L = LcGAN (G,D) + λ1L1 + λ2LSSIM (4) Experimentally the values λ1 = 10 and λ2 = 10 turn out to be good.”
BackgroundMethodInfluentialZhang, R., Wang, Z., Sun, H., Deng, L., & Zhu, H. (2023) · Sensors · doi:10.3390/s23146616 · Cited by 15
“In color fusion (MRI–PET and MRI–SPECT), ChenBlum, MS-SSIM, and SF are lower than other indexes, but remain in the top three.”
MethodSelf-citationDeep Image Prior Amplitude SAR Image Anonymization
Cannas, E. D., Mandelli, S., Bestagini, P., Tubaro, S., & Delp, E. J. (2023) · Remote Sensing · doi:10.3390/rs15153750 · Cited by 11
Mrtnet: Multi-Resolution Temporal Network for Video Sentence Grounding
Ji, W., Qin, Y., Chen, L., Wei, Y., Wu, Y., & Zimmermann, R. (2024) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp48485.2024.10447846 · Cited by 30
“SSIM (Structural Similarity Index Measure) loss [38], is originally devised for image quality assessment.”
BackgroundDeep Bi-directional Attention Network for Image Super-Resolution Quality Assessment
Li, Y., Yang, X., Fu, J., Yue, G., & Zhou, W. M. (2024) · IEEE International Conference on Multimedia and Expo (ICME) · doi:10.1109/icme57554.2024.10687430 · Cited by 37
“These include the multi-scale SSIM (MS-SSIM) [24], the complex wavelet SSIM (CW-SSIM) [23], the gradient magnitude similarity deviation (GMSD) [18], IQA using Kernel Sparse Coding (KSCM) [54], etc.”
MethodSelf-Similarity Measure for Assessment of Image Visual Quality
Ponomarenko, N. N., Jin, L., Лукин, В. В., & Egiazarian, K. (2011) · Lecture notes in computer science · doi:10.1007/978-3-642-23687-7_42 · Cited by 11
“Even the best known metrics as MSSIM [8] possess Spearman rank correlation of this metric and MOS (Mean Opinion Score) of about 0.”
MethodMaking Graphical Information Visible in Real Shadows on Interactive Tabletops
Isogawa, M., Iwai, D., & Sato, K. (2014) · IEEE Transactions on Visualization and Computer Graphics · doi:10.1109/tvcg.2014.2316002 · Cited by 11
Hybrid fractal image coding with quadtree-based progressive structure
Yuen, C.-H., Lui, O.-Y., & Wong, K. (2013) · Journal of Visual Communication and Image Representation · doi:10.1016/j.jvcir.2013.09.002 · Cited by 11
Image distortion analysis and classification scheme using a neural approach
Chetouani, A., Beghdadi, A., & Deriche, M. A. (2010) · European Workshop on Visual Information Processing · doi:10.1109/euvip.2010.5699124 · Cited by 13
An image quality assessment metric based on quaternion wavelet transform
Chen, Q., Xu, Y., Li, C., Liu, N., & Yang, X. (2013) · IEEE International Conference on Multimedia and Expo workshops · doi:10.1109/icmew.2013.6618378 · Cited by 12
“The Correlation Coefficient Value of Different IQA Index IQAc_HPC FSIMc MS-SSIM SSIM awgn 0.8554 0.8758 0.8094 0.8107 awgn-color 0.8993 0.8931 0.8064 0.8029 spatial corr-noise 0.8562 0.8711 0.8195 0.8144 masked noise 0.8123 0.8263 0.8155 0.7795 high-fre-noise 0.9194 0.9115 0.8685 0.8729 impulse noise 0.7365 0.7791 0.6…”
BackgroundMethodInfluentialImage fusion algorithms and metrics duality index
Hossny, M., & Nahavandi, S. (2009) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2009.5413850 · Cited by 11
“Then let x ∈ I be an arbitrary image and 0 ∈ I be a non-informative zero image, e.g. a single colored or a zero contrast image.”
BackgroundGhoreishi, S. E., & Aghvami, A. H. (2016) · IEEE Communications Letters · doi:10.1109/lcomm.2016.2518684 · Cited by 11
“We use the multi-scale structural similarity (MS-SSIM) index [9], which provides a good approximation of userperceived quality.”
MethodLiu, X., Zhang, L., & Lu, K. (2015) · IEEE Transactions on Cloud Computing · doi:10.1109/tcc.2015.2513397 · Cited by 11
Ghlaifan, A., Tounsi, Y., Zada, S., et al. (2016) · Optical Engineering · doi:10.1117/1.oe.55.12.121708 · Cited by 11
Example-based video color transfer
Yao, C.-H., Chang, C.-Y., & Chien, S. (2016) · IEEE International Conference on Multimedia and Expo (ICME) · doi:10.1109/icme.2016.7552926 · Cited by 11
“’s MS-SSIM [6], Pinson et al.”
BackgroundImage decomposition-based structural similarity index for image quality assessment
Yang, J., Lin, Y., Ou, B., & Zhao, X. (2016) · EURASIP Journal on Image and Video Processing · doi:10.1186/s13640-016-0134-5 · Cited by 14
Automatic Detection of Game Engine Artifacts Using Full Reference Image Quality Metrics
Piórkowski, R., Mantiuk, R., & Siekawa, A. (2017) · ACM Transactions on Applied Perception · doi:10.1145/3047407 · Cited by 11
“…Wandell 1997], the texture statistics Structural SIMilarity Index (SSIM) [Wang et al. 2004], the Multiscale Structural SIMilarity Index (MS-SSIM) [Wang et al. 2003], the Color-Image Difference (CID) [Lissner et al. 2013], and the High Dynamic Range Visual Difference Predictor-2 (HDR-VDP-2)…”
BackgroundMethodSAR deception jamming identification via differential feature enhancement
Zhao, B., Huang, L., He, C., et al. (2016) · CIE International Conference on Radar (RADAR) · doi:10.1109/radar.2016.8059174 · Cited by 11
“To show this clearly, the modified SSIM [14]-[17], which evaluates the structural information, luminance, and contrast of the images synthetically, is utilized to measure the changes before and after filtering.”
MethodA new approach for Steganography: Bit shifting operation of encrypted data in LSB (SED-LSB)
Solak, S. B., & Altınışık, U. (2019) · Bilişim Teknolojileri Dergisi · doi:10.17671/gazibtd.435437 · Cited by 11
Fashion Attributes-to-Image Synthesis Using Attention-Based Generative Adversarial Network
Lee, H., & Lee, S. (2019) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv.2019.00055 · Cited by 12
“Lower MS-SSIM score indicates less mode collapse.”
MethodInfluentialQuantitative Reconstruction of Absorption Coefficients for Photoacoustic Tomography
Liu, Y., Sun, M., Liu, T., et al. (2019) · Applied Sciences · doi:10.3390/app9061187 · Cited by 11
Visual Quality Assessment for Motion Compensated Frame Interpolation
Men, H., Lin, H., Hosu, V., Maurer, D., Bruhn, A., & Saupe, D. (2019) · International Conference on Quality of Multimedia Experience (QoMEX) · doi:10.1109/qomex.2019.8743221 · Cited by 9
“There are several FR-IQA methods that consider the HVS, such as SSIM [6], MS-SSIM [7], FSIM [8], VSI [9].”
MethodLearned Quality Enhancement via Multi-Frame Priors for HEVC Compliant Low-Delay Applications
Lu, M., Cheng, M., Xu, Y., et al. (2019) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2019.8803049 · Cited by 11
“For fairness, we follow [3] and evaluate the 3 rd frame of each clip in the Vimeo dataset in terms of PSNR and Struc-( tural SIMilarity (SSIM) [20].”
Identity-Aware Deep Face Hallucination via Adversarial Face Verification
Kazemi, H., Taherkhani, F., & Nasrabadi, N. M. (2019) · IEEE International Conference on Biometrics: Theory, Applications and Systems (BTAS) · doi:10.1109/btas46853.2019.9185984 · Cited by 11
“, textural detail, as it is defined at image pixel level [40, 41].”
BackgroundLiu, J. (2019) · Journal of Visual Communication and Image Representation · doi:10.1016/j.jvcir.2019.102675 · Cited by 11
No‐reference image quality assessment via structural information fluctuation
Yang, X., Wang, T., & Ji, G. (2019) · IET Image Processing · doi:10.1049/iet-ipr.2019.0750 · Cited by 11
“Inspired by the excellent results obtained by SSIM, many variants of SSIM are proposed, such as multi-scale SSIM [13], information content weighted SSIM [14].”
MethodDeep Scalable Image Compression via Hierarchical Feature Decorrelation
Guo, Z., Zhang, Z., & Chen, Z. (2019) · Picture Coding Symposium · doi:10.1109/pcs48520.2019.8954536 · Cited by 11
“If optimized for MS-SSIM (Multiscale Structural Similarity) [20], D denotes the value of 1 - MS-SSIM.”
Methodda Silva, W. B., Mikowski, A., & Casali, R. M. (2020) · IET Image Processing · doi:10.1049/iet-ipr.2019.0941 · Cited by 11
Image Blind Denoising Using a Generative Adversarial Network for LED Chip Visual Localization
Li, D., Gong, S., Niu, S., Wang, Z., Zhou, D., & Lu, H. (2020) · IEEE Sensors Journal · doi:10.1109/jsen.2020.2976576 · Cited by 11
“Because the structural similarity index (SSIM) [36] is able to capture the structural information in an image, we use the multiscale structural similarity index (MS_SSIM) [37] to measure the structural and perceptual similarity between two images.”
MethodSelf-citationSingle Image Blind Deblurring Based on Salient Edge-Structures and Elastic-Net Regularization
Yu, X., & Xie, W. C. (2020) · Journal of Mathematical Imaging and Vision · doi:10.1007/s10851-020-00949-6 · Cited by 13
“RIQA method contacts peak signal-to-noise ratios (PSNR), structural similarity (SSIM), and multi-scale structural similarity (MSSSIM) [36].”
MethodNo-Reference Image Quality Assessment Based on Dual-Domain Feature Fusion
Cui, Y. (2020) · Entropy · doi:10.3390/e22030344 · Cited by 15
Shared Encoder based Denoising of Optical Coherence Tomography Images
Vasudeva, S. A., & Sivaswamy, J. (2018) · Indian Conference on Computer Vision, Graphics and Image Processing · doi:10.1145/3293353.3293388 · Cited by 11
Smarttext: Learning To Generate Harmonious Textual Layout Over Natural Image
Zhang, P., Li, C., & Wang, C. (2020) · IEEE International Conference on Multimedia and Expo (ICME) · doi:10.1109/icme46284.2020.9102780 · Cited by 11
“Accordingly, to obtain more accurate regional segmentation and clear boundaries, we use a hybrid training loss: where L B (Θ) and L S (Θ) denote BCEWithLogits loss [15] and SSIM loss [16], Θ is the BASNet model parameters.”
MethodCapsule GAN for robust face super resolution
Molahasani, M., & Ko, S. (2020) · Multimedia Tools and Applications · doi:10.1007/s11042-020-09489-y · Cited by 10
“Furthermore, as was explained in Section 4.2, MS-SSIM can provide a better approximation for human perception [19].”
BackgroundMethodInfluentialObserver Dependent Lossy Image Compression
Weber, M., Renggli, C., Gräbner, H., & Zhang, C. (2021) · Lecture notes in computer science · doi:10.1007/978-3-030-71278-5_10 · Cited by 22
“Previous research in lossy image compression expressed quality largely in terms of human visual perception and optimized for the human visual system (HVS), using distortion measures such as multiscale structural similarity [47] (MS-SSIM) or mean squared error (MSE) as ar X iv :1 91 0.”
Automated Trace and Copper Plane Extraction of X-ray Tomography Imaged PCBs
Botero, U. J., Ganji, F., Woodard, D. L., & Forte, D. J. (2021) · IEEE Physical Assurance and Inspection of Electronics (PAINE) · doi:10.1109/paine54418.2021.9707715 · Cited by 13
Bombini, A., Anderlini, L., dell'Agnello, L., Giaocmini, F., Ruberto, C., & Taccetti, F. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-06427-2_57 · Cited by 13
Deep-Based Film Grain Removal and Synthesis
Ameur, Z., Hamidouche, W., François, É., Radosavljević, M., Ménard, D., & Demarty, C.-H. (2023) · IEEE Transactions on Image Processing · doi:10.1109/tip.2023.3308726 · Cited by 21
“In all tasks, it has been proved that MS-SSIM helps preserve the contrast in high frequency regions while (cid:96) 1 helps preserve color and luminance, therefore combining them provided relatively better results both in objective and subjective evaluations.”
BackgroundInfluentialTowards anovel perceptual color difference metric using circular processing of hue components
Lee, D., & Plataniotis, K. N. (2014) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp.2014.6853579 · Cited by 13
“0.506 0.382 0.669 0.486 MSSIM [7] 0.775 0.593 0.874 0.691 IDM [13] 0.”
MethodInfluentialImage quality assessment using full-parameter singular value decomposition
Wang, R., Cui, Y.-Z., & Yuan, Y. (2011) · Optical Engineering · doi:10.1117/1.3579459 · Cited by 11
A new approach for image quality assessment using SVD
Azar, F. T., & Amirshahi, S. A. (2007) · International Symposium on Signal Processing and Its Applications (ISSPA) · doi:10.1109/isspa.2007.4555385 · Cited by 11
García, M., Segura-García, J., & Felici-Castell, S. (2017) · Computer Communications · doi:10.1016/j.comcom.2017.08.009 · Cited by 11
Screen content image quality assessment using Euclidean distance
Fu, Y., Zeug, H., Ni, Z., Chen, J., Cai, C., & Ma, K. (2017) · International Symposium on Intelligent Signal Processing and Communication Systems (ISPACS) · doi:10.1109/ispacs.2017.8266443 · Cited by 12
“To investigate effectiveness and superiority, the proposed model is compared with fourteen IQA models, they are PSNR, SSIM [6],MSSIM [17], IWSSIM [18], VIF [10], IFC [11], VSNR [19], MAD [20], FSIM [12], GMSD [8], SPQA [2], GSS [21],SIQM [22].”
MethodA Simple Prediction Fusion Improves Data-driven Full-Reference Video Quality Assessment Models
Bampis, C. G., Bovik, A. C., & Li, Z. (2018) · Picture Coding Symposium · doi:10.1109/pcs.2018.8456293 · Cited by 11
“To demonstrate the potential of E-VMAF, we compared it against leading VQA algorithms, such as PSNR, PSNRhvs [41], SSIM [42], MS-SSIM [43], ST-RRED [14], ST-MAD [10], VQM-VFD [9] and VMAF version 0.”
MethodInfluentialSelf-citationBlind Quality Assessment of 3-D Synthesized Views Based on Hybrid Feature Classes
Rodrigues, F., Ascenso, J., Rodrigues, A., & Queluz, M. P. (2018) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2018.2888830 · Cited by 11
“For scenarios A and AC (cf. Table VI), the PLCC for the proposed metric is again superior (although slightly) to the PLCC for the considered 2D metrics, and with exception for the MS-SSIM metric.”
MethodDeep Image Compression in the Wavelet Transform Domain Based on High Frequency Sub-Band Prediction
Yang, C., Zhao, Y. F., & Wang, S. (2019) · IEEE Access · doi:10.1109/access.2019.2911403 · Cited by 12
“system, mean squared error (MSE) and multi-scale structural similarity index (MS-SSIM) [36] are generally used as the measurement of distortion.”
MethodDeep Convolutional Neural Network with Edge Feature for Image Denoising
Chupraphawan, S., & Ratanamahatana, C. A. (2019) · Advances in intelligent systems and computing · doi:10.1007/978-3-030-19861-9_17 · Cited by 14
Finger-Vein as a Biometric-Based Authentication
Liu, C., Ruan, S.-J., Lai, Y.-R., et al. (2019) · IEEE Consumer Electronics Magazine · doi:10.1109/mce.2019.2941343 · Cited by 11
Özkan, S., & Akar, G. B. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1808.01104 · Cited by 16
“underestimated. Possible solutions for an inverse problem can be exponentially multiplied if only pixel-wise relations are used. Note that this assumption is built for the assessment of image quality [51], [52] and due to its weakness, ME cannot preserve the structural similarity of data. As stated, the loss function…”
MethodDeep Optimization model for Screen Content Image Quality Assessment using Neural NetworksPreprint
Jiang, X., Shen, L., Feng, G., Yu, L., & An, P. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1903.00705 · Cited by 12
Future Frame Prediction Using Convolutional VRNN for Anomaly DetectionPreprint
Lu, Y., Reddy, M. K. K., Nabavi, S. S., et al. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1909.02168 · Cited by 11
“Model Learning: For learning parameters in Conv-VRNN, we combine the least absolute deviation (L1 loss) [24], multi-scale structural similarity measurement (msssim loss) [29] and gradient difference (gdl loss) [21] to define a loss that measure the quality of the predicted frame.”
MethodPezzotti, N., Yousefi, S., Elmahdy, M. S., Gemert, J. H. F. V., Schülke, C., Doneva, M., Tolker‐Nielsen, T., Kastryulin, S., Lelieveldt, B. P. F., van Osch, M. J. P., et al. (2020) · arXiv (Cornell University) · Cited by 13
Image Deconvolution via Noise-Tolerant Self-Supervised InversionPreprint
Kobayashi, H., Solak, A. C., Batson, J., & Royer, L. A. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2006.06156 · Cited by 19
“1 gives averages for four image comparison metrics: Peak Signal to Noise Ratio (PSNR) (Wang Yuanji et al., 2003), Structural Similarity (SSIM) (Wang et al., 2003), Mutual Information (MI) (Russakoff et al., 2004), and Spec- tral Mutual Information (SMI).”
BackgroundInfluentialUnsupervised Image-to-Image Translation via Pre-trained StyleGAN2 NetworkPreprint
Huang, J., Liao, J., & Kwong, S. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2010.05713 · Cited by 11
“We also compared the proposed method with several stateof-the-art work quantitatively, where SSIM [48] was adopted as the matrix to evaluate reconstruction accuracy.”
MethodUnderwater image filtering: methods, datasets and evaluationPreprint
Li, C. Y., Mazzon, R., & Cavallaro, A. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2012.12258 · Cited by 17
“In this section, we discuss the target images used to train neural networks, the chosen network architectures and loss [118], M ULTI - SCALE SSIM (MS-SSIM) [119] OR GRADIENT DIFFERENCE (G RAD -”
MethodA Combined Full-Reference Image Quality Assessment Method Based on Convolutional Activation Maps
Varga, D. (2020) · Algorithms · doi:10.3390/a13120313 · Cited by 11
“437 MS-SSIM [14] 0.”
MethodInfluentialA Near-Field Super-Resolution Network for Accelerating Antenna Characterization
Gu, Y., Sun, H.-H., & van der Weide, D. W. (2024) · IEEE Transactions on Antennas and Propagation · doi:10.1109/tap.2024.3511040 · Cited by 10
Zhang, K.-K., Chen, W.-L., Zhao, T.-S., & Wang, Z. (2025) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2025.3627285 · Cited by 12
“Given the great success of SSIM, several its variants have been proposed, such as multi-scale SSIM [20] and information content weighted SSIM [21].”
Self-citationConsistency-Aware Map Generation at Multiple Zoom Levels Using Aerial Image
Chen, L., Fang, Z., & Fu, Y. (2026) · IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing · doi:10.1109/jstars.2022.3170591 · Cited by 12
Underwater Image Restoration with Light-Aware Progressive Network
Yang, J., Li, C., & Li, X. (2023) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49357.2023.10095380 · Cited by 11
No-Reference Image Quality Assessment Based on a Multitask Image Restoration Network
Chen, F., Fu, H. C., Yu, H., & Chu, Y. (2023) · Applied Sciences · doi:10.3390/app13116802 · Cited by 14
Learned Distributed Image Compression with Multi-Scale Patch Matching in Feature Domain
Huang, Y., Chen, B., Qin, S., Li, J., Wang, Y., Dai, T., & Xia, S. (2023) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v37i4.25551 · Cited by 13
Konx: cross-resolution image quality assessment
Wiedemann, O., Hosu, V., Su, S., & Saupe, D. (2023) · Quality and User Experience · doi:10.1007/s41233-023-00061-8 · Cited by 12
“Some works on full-reference IQA [25, 26] integrate information from downscaled versions of their input internally.”
Enlighten Fusion Multiscale Network for Infrared and Visible Image Fusion in Dark Environments
Wang, H., Shu, C., & Li, X. (2023) · IEEE Signal Processing Letters · doi:10.1109/lsp.2023.3309153 · Cited by 10
“For the MS-SSIM metric, the performance of EFMN is excusable since the overall brightness and contrast has changed a lot compared with that of the original low-light image.”
MethodHigh Visual-Fidelity Learned Video Compression
Li, M., Shi, Y., Wang, J., & Huang, Y. (2023) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3581783.3612530 · Cited by 21
“These methods often incorporate perceptual loss functions, such as structural similarity (SSIM) [27] or multi-scale structural similarity (MS-SSIM) [28], LPIPS [33], to guide the compression process and improve the visual quality of the reconstructed videos.”
MethodGenerating a full spherical view by modeling the relation between two fisheye images
Flores, M., Valiente, D., Peidró, A., et al. (2024) · The Visual Computer · doi:10.1007/s00371-024-03293-7 · Cited by 10
“In addition, an extended version called Multi-Scale Structural Similarity Index Method (MS-SSIM) [47] evaluates the structural similarity of both images at different image scales.”
Faithful Model Explanations through Energy-Constrained Conformal Counterfactuals
Altmeyer, P., Farmanbar, M., Arie Van Deursen, A., & Liem, C. C. S. (2024) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v38i10.28956 · Cited by 14
Compressed-Domain Vision Transformer for Image Classification
Ji, R., & Karam, L. J. (2024) · IEEE Journal on Emerging and Selected Topics in Circuits and Systems · doi:10.1109/jetcas.2024.3394878 · Cited by 12
“In this paper, we employ the variational image compression with a scale hyperprior [28] using the MS-SSIM [35] quality metric in the loss function (HyperMS-SSIM).”
MethodNejat, F., Eghtedari, S., & Alimoradi, F. (2024) · Ophthalmology Science · doi:10.1016/j.xops.2024.100546 · Cited by 11
“The third loss function is Structural Similarity Index loss 24 which helps the model to learn structural information.”
BackgroundSurface defect detection and semantic segmentation with a novel lightweight deep neural network
Huang, Q., Li, F., Yang, Y., Tao, X., Li, W., Xu, W., & Wang, Y. (2024) · Measurement Science and Technology · doi:10.1088/1361-6501/ad4ab2 · Cited by 11
“In the binary defect segmentation tasks, our loss function is composed of the sum of binary cross-entropy (BCE) loss [37], structural similarity (SSIM) loss [38], boundary IoU loss [39], and F-measure loss [40].”
An Image Dehazing Algorithm for Underground Coal Mines Based on gUNet
Tian, F., Gao, L., & Zhang, J. (2024) · Sensors · doi:10.3390/s24113422 · Cited by 10
Enhancing a Display’s Sunlight Readability with Tone Mapping
Qian, Y., Chen, S.-C., Hsiang, E., Akimoto, H., Lin, C., & Wu, S.-T. T. (2024) · Photonics · doi:10.3390/photonics11060578 · Cited by 10
IV-SSIM—The Structural Similarity Metric for Immersive Video
Dziembowski, A., Nowak, W., & Stankowski, J. (2024) · Applied Sciences · doi:10.3390/app14167090 · Cited by 12
Attention-oriented residual block for real-time low-light image enhancement in smart ports
Li, L., Xu, W., Gao, Y., Lu, Y., Yang, D., Liu, R. W., & Zhang, R. (2024) · Computers & Electrical Engineering · doi:10.1016/j.compeleceng.2024.109634 · Cited by 12
Infrared Image Generation Based on Visual State Space and Contrastive Learning
Li, B. K., Ma, D., He, F., et al. (2024) · Remote Sensing · doi:10.3390/rs16203817 · Cited by 10
Advancing dermoscopy through a synthetic hair benchmark dataset and deep learning-based hair removal
Jütte, L., Patel, H., & Roth, B. (2024) · Journal of Biomedical Optics · doi:10.1117/1.jbo.29.11.116003 · Cited by 10
Shen, G., Ouyang, Y., Lu, J., Yang, Y., & Sánchez, V. (2024) · IEEE Transactions on Image Processing · doi:10.1109/tip.2024.3512369 · Cited by 16
“In practice, SSIM is locally calculated by sliding a Gaussian weighing filter over the entire image [79], [80].”
Personalized Image Generation with Large Multimodal Models
Xu, Y., Wang, W., Zhang, Y., Tang, B., Yan, P., Feng, F., & He, X. (2025) · International World Wide Web Conference (WWW) · doi:10.1145/3696410.3714843 · Cited by 50
Generative Latent Coding for Ultra-Low Bitrate Image and Video Compression
Qi, L., Jia, Z., Li, J., Li, B., Li, H., & Lu, Y. (2025) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2025.3571944 · Cited by 63
“Most existing learned video compression methods are designed to optimize traditional metrics like PSNR or MS-SSIM [42].”
DocScanner: Robust Document Image Rectification with Progressive Learning
Feng, H., Zhou, W., Deng, J., Tian, Q. C., & Li, H. (2025) · International Journal of Computer Vision · doi:10.1007/s11263-025-02431-5 · Cited by 40
“Third, we use Multi-Scale Structural Similarity (MS-SSIM) (Wang et al., 2003) to measure the image similarity between the rectified images and the ground truth undistorted images.”
InfluentialSMC++: Masked Learning of Unsupervised Video Semantic Compression
Tian, Y., Ling, X., Geng, C., Hu, Q., Lu, G., & Zhai, G. (2025) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2025.3625063 · Cited by 33
“Video Compression: Previous video codecs, including traditional ones [1], [2], [36] and learnable ones [3], [4], [37], [38], [39], [40], are designed to achieve better pixel-wise signal quality metrics, e.g., PSNR and MS-SSIM [41], which mainly serve the human visual experience.”
Perceptual evaluation of image denoising algorithms
Zeng, K., & Wang, Z. (2013) · Asilomar Conference on Signals, Systems and Computers (ACSSC) · doi:10.1109/acssc.2013.6810514 · Cited by 11
“150 MS-SSIM[22] FR 0.”
Self-citationPower‐efficient real‐time solution for adaptive vision algorithms
Tabkhi, H., Sabbagh, M., & Schirner, G. (2014) · IET Computers & Digital Techniques · doi:10.1049/iet-cdt.2014.0075 · Cited by 10
Machine learning based modeling of spatial and temporal factors for video quality assessment
Narwaria, M., & Lin, W. (2011) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2011.6116173 · Cited by 11
“Let SQi denote the spatial quality of the ith frame using MS-SSIM.”
MethodSequential Error Concealment for Video/Images by Weighted Template Matching
Koloda, J., Østergaard, J. S., Jensen, S. H., et al. (2012) · DCC · doi:10.1109/dcc.2012.24 · Cited by 10
Fast protection of H.264/AVC by reduced selective encryption of CAVLC
Dubois, L., Puech, W., & Blanc-Talon, J. (2011) · HAL (Le Centre pour la Communication Scientifique Directe) · Cited by 10
Flicker sensitive motion tuned video quality assessment
Choi, L. K., & Bovik, A. C. (2016) · IEEE Southwest Symposium on Image Analysis and Interpretation (SSIAI) · doi:10.1109/ssiai.2016.7459167 · Cited by 18
“PSNR, MS-SSIM, and VSNR were applied on a frame-by-frame basis and the average score across all frames were used as a final quality score.”
MethodSelf-citationReduced-reference image quality assessment based on DCT Subband Similarity
Balanov, A., Schwartz, A., & Moshe, Y. (2016) · International Conference on Quality of Multimedia Experience (QoMEX) · doi:10.1109/qomex.2016.7498930 · Cited by 11
“Four FR IQA metrics - PSNR, SSIM [2], MS-SSIM [3] and DSS [4], and four RR IQA metrics – WNSIM [9], RRED [13], RDCT [18] and RD-DSS, were examined.”
MethodInfluentialLi, Z., Ramos, A., Li, M., et al. (2020) · Biomedical Physics & Engineering Express · doi:10.1088/2057-1976/ab8fc6 · Cited by 10
A Training Method for Image Compression Networks to Improve Perceptual Quality of Reconstructions
Lee, J., Kim, D., Kim, Y., et al. (2020) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw50498.2020.00080 · Cited by 10
“…efficiency and the latest approach [15] has obtained a remarkable coding efficiency that outperforms Ver-satile Video Coding (VVC) Intra (VTM 7.1 [1]), which has been almost finalized for standardization by ISO/IEC MPEG, in terms of both PSNR and multi-scale structural similarity index (MS-SSIM) [25].”
MethodInfluentialLearning Surrogates via Deep Embedding
Patel, Y., Hodaň, T., & Matas, J. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-58577-8_13 · Cited by 18
“For example, the state-of-the-art learned image compression [33,4] and super-resolution [32,9] methods directly optimize the perceptual similarity metrics such as MS-SSIM [59] and the peak signal-to-noise ratio (PSNR).”
MethodPredicting Spatio-Temporal Entropic Differences for Robust No Reference Video Quality Assessment
Mitra, S., Soundararajan, R., & Channappayya, S. S. (2021) · IEEE Signal Processing Letters · doi:10.1109/lsp.2021.3049682 · Cited by 9
“A popular approach for designing NR VQA algorithms is to use pseudo-labels in the form of frame level FR image quality assessment (QA) measures such as SSIM [8] or MS-SSIM [9].”
Method
Metadata
- Authors
- Wang, Z., Simoncelli, E. P., & Bovik, A. C.
- Affiliations
- Courant Institute of Mathematical Sciences (US)
- The University of Texas at Austin (US)
- Venue
- The Thrity-Seventh Asilomar Conference on Signals, Systems & Computers, 2003
- Publisher
- IEEE
- Published
- 2004-07-08
- Pages
- 1398-1402
- ISSN
- 1058-6393 (ISSN-L)
- OpenAlex
- W1580389772
- Cited by
- 6,253
- References
- 15
- Citations by year
- 2026: 344 · 2025: 675 · 2024: 729 · 2023: 668 · 2022: 662 · 2021: 641 · 2020: 563 · 2019: 438 · 2018: 305 · 2017: 267 · 2016: 206 · 2015: 173 · 2014: 166 · 2013: 112 · 2012: 89
- FWCI
- 3.98
- Type
- conference paper
- Language
- en
- Open access
- closed
- Locations
- Asilomar Conference on Signals, Systems and Computers (ACSSC) (publishedVersion)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Image and Video Quality Assessment
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Image Retrieval and Classification Techniques
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Digital Media Forensic Detection
- Keywords
- structural similarity, image quality assessment, multiscale analysis, human visual system, perceptual image quality, image generation, parameter calibration, viewing conditions
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.