Multiscale structural similarity for image quality assessment
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.
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. -
More than one paper fits this citation.
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 · cited 1×
“Traditional perceptual image quality assessment methods are based on a bottom-up approach which attempts to simulate the functionality of the relevant early human visual system (HVS) components.”
BackgroundA 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 · cited 2×
“These synthesis methods provide a powerful and efficient means of testing a model, and have the added benefit that the resulting images suggest improvements that might be made to the model [11].”
MethodSelf-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 919
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 801–900 of 6,047
Structural Prior Driven Regularized Deep Learning for Sonar Image Classification
Gerg, I. D., & Monga, V. (2021) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2020.3045649 · Cited by 50
“To encourage image enhancement for classification, we utilize a novel loss function between the desired enhanced image and the original, the multiscale structural similarity measure (MS-SSIM) [47], [48] which is a scale-aware version of the SSIM measure”
MethodModulated Contrast for Versatile Image Synthesis
Zhan, F., Zhang, J., Yu, Y., Wu, R., & Lu, S. (2022) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52688.2022.01774 · Cited by 75
“Among various distance metrics [36, 37], perceptual loss [17] emerges as a powerful metric in line with human perception by leveraging the internal activation of pre-trained networks.”
BackgroundLow-rate Image Compression with Super-resolution Learning
Gao, W., Tao, L., Zhou, L., et al. (2020) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw50498.2020.00085 · Cited by 47
“(9) by combining mean squared error (MSE) and multi-scale structural similarity (MS-SSIM) [19] score, L = Dw +Rû +Rv̂ (8) Dw = λ1 × ‖x− x̂‖22 + λ2 × (1−MS-SSIM) (9) where λ1 and λ2 denote weighting coefficients, x and x̂ denote original and compressed images, respectively.”
MethodInfluentialMadhusudana, P. C., Birkbeck, N., Wang, Y., Adsumilli, B., & Bovik, A. C. (2021) · IEEE Transactions on Image Processing · doi:10.1109/tip.2021.3106801 · Cited by 63
“Since our proposed framework is an FR/RR model, we selected 4 popular and widely used FRIQA methods: PSNR, SSIM [9], MS-SSIM [10] and FSIM [11] for comparison.”
MethodSelf-citationLin, D., Li, Y., Prasad, S., Nwe, T. L., Dong, S., & Oo, Z. M. (2021) · Knowledge-Based Systems · doi:10.1016/j.knosys.2021.107272 · Cited by 51
Dense Semantics-Assisted Networks for Video Action Recognition
Luo, H., Lin, G., Yao, Y., Tang, Z., Wu, Q., & Hua, X. (2021) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2021.3100842 · Cited by 55
Ost, J., Laradji, I., Newell, A., Bahat, Y., & Heide, F. (2022) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52688.2022.01787 · Cited by 52
“For PSNR and SSIM, higher is better; for LPIPS lower is better.”
MethodComparison of single image HDR reconstruction methods — the caveats of quality assessment
Hanji, P., Mantiuk, R., Eilertsen, G., Hajisharif, S., & Unger, J. (2022) · ACM SIGGRAPH Conference (SIGGRAPH) · doi:10.1145/3528233.3530729 · Cited by 65
Two-Stage Edge Reuse Network for Salient Object Detection of Strip Steel Surface Defects
Han, C., Li, G., & Liu, Z. (2022) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2022.3200114 · Cited by 53
Al‐Huda, Z., Peng, B., Algburi, R. N. A., Alfasly, S., & Li, T. (2022) · Applied Intelligence · doi:10.1007/s10489-022-04212-w · Cited by 49
“Besides, the SSIM loss LSSIM [48] is developed to accurately measure cracks of various widths by predicting the crack structure in terms of spatial location.”
MethodZhang, Z., Sahu, C. K., Singh, S., Rai, R., Yang, Z., & Lu, Y. (2023) · International Journal of Production Research · doi:10.1080/00207543.2023.2201860 · Cited by 53
Transformer-Based Cross-Modal Integration Network for RGB-T Salient Object Detection
Lv, C., Zhou, X., Wan, B., Wang, S., Sun, Y., Zhang, J., & Yan, C. C. (2024) · IEEE Transactions on Consumer Electronics · doi:10.1109/tce.2024.3390841 · Cited by 58
“In addition, we add the SSIM loss [68] to compute the loss value of our network, where we employ the BCE + IoU + SSIM loss to supervise the network training ( w SSIM).”
MethodLogo Synthesis and Manipulation with Clustered Generative Adversarial Networks
Sage, A., Timofte, R., Agustsson, E., & Gool, L. V. (2018) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2018.00616 · Cited by 88
“Lower values mean higher quality for CORNIA and higher diversity for MS-SSIM. above 32 clusters.”
MethodInfluentialVasudeva, S. A., & Sivaswamy, J. (2019) · The Springer series on challenges in machine learning · doi:10.1007/978-3-030-25614-2_4 · Cited by 55
“this purpose. The SSIM is measured at a xed scale and may only be appropriate for a certain range of image scales. A more advanced form of SSIM is the multi-scale structure similarity index (MS-SSIM) [18]. It preserves the structure and contrast in high-frequency regions better than the other loss functions [21]. In a…”
BackgroundConvolutional Neural Networks for Continuous QoE Prediction in Video Streaming Services
Duc, T. N., Tran, C. M., Tan, P. X., & Kamioka, E. (2020) · IEEE Access · doi:10.1109/access.2020.3004125 · Cited by 53
“Reduced Reference Entropic Differences (STRRED) [33], Multi-Scale Structural Similarity (MS-SSIM) [34], Peak Signal to Noise Ratio (PSNR) [35], etc.”
Babu, K. K., & Dubey, S. R. (2020) · Neurocomputing · doi:10.1016/j.neucom.2020.06.104 · Cited by 47
Activation Maximization Generative Adversarial NetsPreprint
Zhou, Z., Cai, H., Rong, S., Song, Y., Ren, K., Zhang, W., Yu, Y. H., & Wang, Y.-X. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1703.02000 · Cited by 82
Distribution Matching in Variational Inference
Rosca, M., Lakshminarayanan, B., & Mohamed, S. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1802.06847 · Cited by 42
“7: Update generator θ by minimizing (cid:46) data reconstruction and generation loss 8: Update discriminator φ by minimizing treat data as real, reconstructions and samples as fake 9: Update code discriminator ω by minimizing treat p ( z ) as real and codes from the encoder as fake 10: end for use 1.0 - MS-SSIM (Wang…”
MethodInfluentialSushko, V. (2024) · bonndoc (University of Bonn) · Cited by 50
Objective Quality Assessment of Lenslet Light Field Image Based on Focus Stack
Meng, C., An, P., Huang, X., Yang, C., Shen, L., & Wang, B. (2021) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2021.3096071 · Cited by 47
PU-M-Net for phase unwrapping with speckle reduction and structure protection in ESPI
Xu, M. G., Tang, C., Shen, Y., et al. (2021) · Optics and Lasers in Engineering · doi:10.1016/j.optlaseng.2021.106824 · Cited by 46
Beauferris, Y., Teuwen, J., Karkalousos, D., Moriakov, N., Caan, M. W. A., Yiasemis, G., Rodrigues, L., Lopes, A., Pedrini, H., Rittner, L., et al. (2022) · Frontiers in Neuroscience · doi:10.3389/fnins.2022.919186 · Cited by 62
“They used a linear combination of VIF andMS-SSIM (Wang et al., 2003) as their loss function.”
MethodOpen Set Recognition With Incremental Learning for SAR Target Classification
Ma, X., Ji, K., Feng, S., Zhang, L., Xiong, B., & Kuang, G. (2023) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2023.3283423 · Cited by 52
“Studies have shown that MS-SSIM loss has better generalization ability [38].”
BackgroundMethodDigital staining in optical microscopy using deep learning - a review
Kreiss, L. A., Jiang, S., Li, X., Xu, S., Zhou, K. C., Lee, K. C., Mühlberg, A., Kim, K., Chaware, A., Ando, D. M., et al. (2023) · PhotoniX · doi:10.1186/s43074-023-00113-4 · Cited by 54
“The innovations in the fields of convolutional networks, the U-Net and generative models, were accompanied by a series of new quantitative metrics for image similarity, such as Wasserstein loss [131], structural similarity index (SSIM) [132] or multi-scale SSIM (MS-SSIM) [133].”
Methodcomprehensive review of image super-resolution metrics: classical and AI-based approaches
Ogli, A. M. M., Ogli, B. S. A., Rikhsivoev, M., Nosirov, K., & Saydiakbarov, S. (2024) · ACTA IMEKO · doi:10.21014/actaimeko.v13i1.1679 · Cited by 78
Structure and Hue Similarity for Color Image Quality Assessment
Shi, Y., Ding, Y., Zhang, R., & Li, J. (2009) · International Conference on Electronics Computer Technology · doi:10.1109/icect.2009.116 · Cited by 55
Sequential Error Concealment for Video/Images by Sparse Linear Prediction
Koloda, J., Østergaard, J., Jensen, S. H., et al. (2013) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2013.2238524 · Cited by 46
A Saliency Dispersion Measure for Improving Saliency-Based Image Quality Metrics
Zhang, W., Martin, R. R., & Liu, H.-T. (2018) · IEEE transactions on circuits and systems for video technology (Print) · doi:10.1109/tcsvt.2017.2650910 · Cited by 36
“metrics, including PSNR [1], UQI [18], SSIM [19], MS-SSIM [20], VIF [21], and FSIM [22]; and four no-reference (NR) metrics,”
MethodWang, J., Wang, S., & Wang, Z. (2017) · IEEE Transactions on Image Processing · doi:10.1109/tip.2017.2651387 · Cited by 50
“scale SSIM (MS-SSIM) [8], universal quality index (UQI) [9], and visual information fidelity (VIF) [10] measurements of left- and right-views performs equally well or better than the advanced 3D-IQA or 3D-VQA models [11]–[19] on LIVE 3D Image Database Phase I [2].”
MethodSelf-citationCVIQD: Subjective quality evaluation of compressed virtual reality images
Sun, W., Gu, K., Zhai, G., Ma, S., Lin, W., & Calle, P. L. (2017) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2017.8296923 · Cited by 47
“Some classical IQA methods [17, 18] were therefore proposed mainly depending on structural information or statistic information.”
MethodInfluentialZhu, W., Zhai, G., Min, X., Hu, M., Liu, J., Guo, G., & Yang, X. (2019) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2019.2902484 · Cited by 51
Xu, L., Zeng, X., Zhang, H., Li, W., Lei, J., & Huang, Z. (2020) · Neural Networks · doi:10.1016/j.neunet.2020.05.001 · Cited by 46
Qu, Q., Chen, X., Chung, Y. Y., & Chen, Z. (2021) · IEEE Transactions on Broadcasting · doi:10.1109/tbc.2021.3099737 · Cited by 50
“2D IQA is a well-studied problem with numerous 2D FR-IQA metrics proposed such as SSIM, [8], multi-scale SSIM [13], FSIM [14], IWSSIM [15], visual saliency-induced index (VSI) [16], gradient magnitude similarity deviation (GMSD) [17], efficient Minkowski distance-based metric [18], multiple pseudo…”
Perceptual Quality Assessment of HEVC and VVC Standards for 8K Video
Bonnineau, C., Hamidouche, W., Fournier, J., et al. (2022) · IEEE Transactions on Broadcasting · doi:10.1109/tbc.2022.3140710 · Cited by 46
“In this experiment, objective quality metrics, including PSNR, multi-scale structural similarity (MS-SSIM) [28], and VMAF [17], are used to measure the distortion between the 8K reconstructed signal and the source video.”
MethodInfluentialMooers, G., Pritchard, M., Beucler, T., Ott, J., Yacalis, G., Baldi, P., & Gentine, P. (2021) · Journal of Advances in Modeling Earth Systems · doi:10.1029/2020ms002385 · Cited by 90
“To better quantify the differences between true and predicted PSDs, we rely on the Log-Spectral Distance (D) (Wang et al., 2003):”
MethodHigh-Performance Polarization Remote Sensing With the Modified U-Net Based Deep-Learning Network
Li, D., Lin, B., Wang, X., & Guo, Z. (2022) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2022.3164917 · Cited by 51
“Peak signal-to-noise ratio (PSNR) and structural similarity index (SSIM) are commonly used as indicators for image quality evaluation [60].”
MethodVirtual Stain Transfer in Histology via Cascaded Deep Neural Networks
Yang, X., Bai, B., Zhang, Y., Li, Y., Haan, K. D., Liu, T., & Özcan, A. (2022) · ACS Photonics · doi:10.1021/acsphotonics.2c00932 · Cited by 52
Toward Robust Neural Image Compression: Adversarial Attack and Model Finetuning
Chen, T., & Ma, Z. (2023) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2023.3276442 · Cited by 64
“19 in the supplemental material that other distance metrics, such as the l 1 distance and MS-SSIM [15], can be used to guide the example generation.”
MethodGAN-based generation of realistic 3D volumetric data: A systematic review and taxonomy
Ferreira, A. F., Li, J., Pomykala, K. L., Kleesiek, J., Alves, V., & Egger, J. (2024) · Medical Image Analysis · doi:10.1016/j.media.2024.103100 · Cited by 88
AnlightenDiff: Anchoring Diffusion Probabilistic Model on Low Light Image Enhancement
Chan, C.-Y., Siu, W.-C., Chan, Y., & Chan, H. A. (2024) · IEEE Transactions on Image Processing · doi:10.1109/tip.2024.3486610 · Cited by 68
“R EFERENCE C OMPARISON ON LOL [5], VELOL [47] AND LOL V 2 [48] D ATASETS IN T ERMS OF PSNR, SSIM [50], AND LPIPS [19].”
MethodSurvey of information theory in visual quality assessment
Soundararajan, R., & Bovik, A. C. (2013) · Signal Image and Video Processing · doi:10.1007/s11760-013-0442-5 · Cited by 50
“Then, the weighted VQA index is given by Indeed, it is verified in [35] that the performance of weighting the SSIM index based on perceptual uncertainty of motion does improve its performance over the algorithm that does not perform such a weighting.”
MethodInfluentialSelf-citationDisplay adaptive 3D content remapping
Masiá, B., Wetzstein, G., Aliaga, C., Raskar, R., & Gutiérrez, D. (2013) · Computers & Graphics · doi:10.1016/j.cag.2013.06.004 · Cited by 48
Decision Fusion for Image Quality Assessment using an Optimization Approach
Oszust, M. (2016) · IEEE Signal Processing Letters · doi:10.1109/lsp.2015.2500819 · Cited by 37
Lv, Y., Jiang, G., Yu, M., Xu, H., Shao, F., & Liu, S. (2015) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2015.7351221 · Cited by 50
Tamboli, R. R., Appina, B., Channappayya, S. S., & Jana, S. (2016) · Signal Processing Image Communication · doi:10.1016/j.image.2016.05.010 · Cited by 48
Light Filed Image Quality Assessment by Local and Global Features of Epipolar Plane Image
Fang, Y., Wei, K., Hou, J., Wen, W., & İmamoğlu, N. (2018) · IEEE International Conference on Multimedia Big Data · doi:10.1109/bigmm.2018.8499086 · Cited by 51
A Distribution and Structure Match Generative Adversarial Network for SAR Image Classification
Ren, Z., Hou, B., Wu, Q., Wen, Z., & Jiao, L. (2020) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2019.2959120 · Cited by 47
“Limited to the size of patches, we do not set the same scales of MS-SSIM as [59] in our experiments.”
MethodInfluentialJiang, Q., Zhou, W., Chai, X., Yue, G., Shao, F., & Chen, Z. (2020) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2020.3005111 · Cited by 49
“The competing FR 2-D-IQM methods include SSIM [7], MSSSIM [8], VIF [9], MAD [10], and DeepSIM [40].”
MethodWang, R., Song, P., Jiang, S., Yan, C., Zhu, J., Guo, C., Bian, Z., Wang, T., & Zheng, G. (2020) · Optics Letters · doi:10.1364/ol.400244 · Cited by 45
Singh, P., Diwakar, M., Cheng, X., & Shankar, A. (2021) · Journal of Real-Time Image Processing · doi:10.1007/s11554-021-01125-8 · Cited by 44
“The similarity between LH1 and LH11 is calculated based on features like luminance, contrast, and structure [43, 44] using Eq.”
BackgroundMethodDeep learning-based statistical noise reduction for multidimensional spectral data
Kim, Y., Oh, D., Huh, S., Song, D., Jeong, S., Kwon, J., Kim, M., Kim, D., Ryu, H., Jung, J., et al. (2021) · Review of Scientific Instruments · doi:10.1063/5.0054920 · Cited by 52
“The loss function is defined as the weighted sum of mean absolute error and multiscale structural similarity 23,24 .”
Choi, J. W., Cho, Y. J., Ha, J. Y., Lee, S. B., Lee, S., Choi, Y. H., Cheon, J., & Kim, W. S. (2021) · Scientific Reports · doi:10.1038/s41598-021-00058-3 · Cited by 66
“For a CT volume with a shape of 512 × 512 × N, running the generator network through the entire CT volume would first yield N-15 arrays with a shape of 512 × 512 × 16 that partially overlap with one another.”
BackgroundMethodInfluentialImage utility assessment and a relationship with image quality assessment
Rouse, D., Pépion, R., Hemami, S. S., & Callet, P. L. (2009) · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · doi:10.1117/12.811664 · Cited by 50
Video Compression With CNN-Based Postprocessing
Zhang, F., Ma, D., Chen, F., & Bull, D. (2021) · IEEE Multimedia · doi:10.1109/mmul.2021.3052437 · Cited by 43
“In the first stage, the generator is trained separately using the multi-scale structural similarity index (MS-SSIM) [18] as the loss function.”
MethodS3IM: Stochastic Structural SIMilarity and Its Unreasonable Effectiveness for Neural Fields
Xie, Z., Yang, X.-D., Yang, Y., Sun, Q., Jiang, Y.-X., Wang, H., Cai, Y., & Sun, M. (2023) · IEEE International Conference on Computer Vision · doi:10.1109/iccv51070.2023.01652 · Cited by 64
“In image quality assessment, the point-wise MSE-based metric Peak Signal-to-Noise Ratio (PSNR) [7] is widely used but do not correlate well with perceived image quality [26, 41], since point-wise metrics do not reflect structural information [16, 42, 43, 32] that is rich in the physical world.”
BackgroundNguyen, L. X., Tun, Y. L., Tun, Y. K., Nguyen, M. N. H., Zhang, C., Han, Z., & Hong, C. S. (2024) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2024.3362328 · Cited by 57
“Therefore, [34] introduced multi-scale SSIM to incorporate image details at different resolutions by iteratively applying a low-pass filter and downsampling by a factor of two.”
BackgroundMethodFFTFormer: A spatial-frequency noise aware CNN-Transformer for low light image enhancement
Pei, X., Huang, Y., Su, W., Zhu, F., & Liu, Q. (2025) · Knowledge-Based Systems · doi:10.1016/j.knosys.2025.113055 · Cited by 53
Sparse Representation-Based Image Quality Index With Adaptive Sub-Dictionaries
Li, L.-D., Cai, H., Zhang, Y.-B., Lin, W., Kot, A., & Sun, X.-M. (2016) · IEEE Transactions on Image Processing · doi:10.1109/tip.2016.2577891 · Cited by 46
“A multi-scale version, i.e., MS-SSIM, was later developed to account for the influence of varying viewing conditions on the perceived quality [20].”
MethodLearning Blind Quality Evaluator for Stereoscopic Images Using Joint Sparse Representation
Shao, F., Li, K., Lin, W., Jiang, G., & Dai, Q. (2016) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2016.2594142 · Cited by 44
LaBGen: A method based on motion detection for generating the background of a scene
Laugraud, B., Piérard, S. L., & Droogenbroeck, M. V. (2016) · Pattern Recognition Letters · doi:10.1016/j.patrec.2016.11.022 · Cited by 44
“Multi-Scale Structural Similarity Index (MS-SSIM, ↑, [−1, 1]): metric defined by Wang et al. (2003) using structural distortion as an estimation of the perceived visual distortion.”
MethodFrom Sparse Coding Significance to Perceptual Quality: A New Approach for Image Quality Assessment
Ahar, A., Barri, A., & Schelkens, P. (2018) · IEEE Transactions on Image Processing · doi:10.1109/tip.2017.2771412 · Cited by 35
Color-Sensitivity-Based Combined PSNR for Objective Video Quality Assessment
Shang, X., Liang, J., Guo-zhong, W., Zhao, H., Wu, C., & Lin, C. (2018) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2018.2836974 · Cited by 50
“Many methods, such as the structure similarity (SSIM) [6]–[8], video quality metric (VQM) [9], visual signal-to-noise ratio (VSNR) [10], visual information fidelity (VIF) [11], motion-based video integrity evaluation index (MOVIE) [12], feature similarity (FSIM) index [13], and works in [14]–[16], have been proposed t…”
MethodReconstruction Distortion Oriented Light Field Image Dataset for Visual Communication
Huang, Z., Yu, M., Jiang, G., et al. (2019) · International Symposium on Networks, Computers and Communications (ISNCC) · doi:10.1109/isncc.2019.8909170 · Cited by 47
“Based on the consideration of spatial domain distortion, the full reference 2D image metrics PSNR, SSIM [18], MS-SSIM [19], UQI [20], VIF [21], VSI [22] and GMSD [23] are selected.”
MethodCNN-based superresolution reconstruction of 3D MR images using thick-slice scans
Jurek, J., Kociński, M., Materka, A., Elgalal, M. T., & Majos, A. (2019) · Journal of Applied Biomedicine · doi:10.1016/j.bbe.2019.10.003 · Cited by 49
Jiang, Q., Gao, W., Wang, S., Yue, G., Shao, F., Ho, Y., & Kwong, S. (2020) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2020.2984928 · Cited by 45
Duan, C., Pan, J. H., & Li, R. (2020) · Remote Sensing · doi:10.3390/rs12203446 · Cited by 47
Subjective and Objective Quality Assessment of Compressed Screen Content Videos
Li, T., Min, X., Zhao, H., Zhai, G., Xu, Y., & Zhang, W. (2020) · IEEE Transactions on Broadcasting · doi:10.1109/tbc.2020.3028335 · Cited by 45
“Based on SSIM, a bunch of modification models such as Multi-Scale SSIM (MS-SSIM) [21] and Information Weighted SSIM (IW-SSIM) [22], assume that HVS tends to perceive the local structures in a multi-scale manner.”
MethodSynthesis of 3D MRI Brain Images With Shape and Texture Generative Adversarial Deep Neural Networks
Chong, C. K., & Ho, E. T. W. (2021) · IEEE Access · doi:10.1109/access.2021.3075608 · Cited by 49
“To compare the diversity of synthetic image stacks against the diversity present in the real training cohort, we computed the Multi-Scale Structural Similarity Index (MS-SSIM) [41].”
MethodImage Super-Resolution Quality Assessment: Structural Fidelity Versus Statistical Naturalness
Zhou, W., Wang, Z., & Chen, Z. (2021) · International Conference on Quality of Multimedia Experience (QoMEX) · doi:10.1109/qomex51781.2021.9465479 · Cited by 69
“Given the original HR image x , and the SISR generated test image y , inspired by the success of MS-SSIM [17], we define subband patch level structural fidelity measure as: where x and y denote the patches extracted from the k -th subband from x and y , respectively, σ x and σ y are their standard…”
InfluentialSelf-citationBlind Stereoscopic Image Quality Evaluator Based on Binocular Semantic and Quality Channels
Sim, K., Yang, J., Lu, W. B., & Gao, X. (2021) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2021.3064240 · Cited by 47
Spatiotemporal Representation Learning for Blind Video Quality Assessment
Liu, Y., Wu, J., Li, L., Dong, W., Zhang, J., & Shi, G. (2021) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2021.3114509 · Cited by 48
Learned Image Compression with Discretized Gaussian Mixture Likelihoods and Attention ModulesPreprint
Cheng, Z., Sun, H., Takeuchi, M., et al. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2001.01568 · Cited by 44
“To achieve high subjective quality, we also train the model using MS-SSIM quality metrics [33] and then distortion term is defined by D ( x , ˆ x ) = 1 − MS - SSIM( x , ˆ x ) .”
MethodPractical Evaluation of VMAF Perceptual Video Quality for WebRTC Applications
García, B., López-Fernández, L., Gortázar, F., & Gallego, M. (2019) · Electronics · doi:10.3390/electronics8080854 · Cited by 44
“• MS-SSIM: Multi-Scale SSIM, which is an advanced form of SSIM in which the evaluation is conducted through a process of sub-sampling, reminiscent of multiscale processing [27].”
MethodApplicability of Existing Objective Metrics of Perceptual Quality for Adaptive Video Streaming
Søgaard, J., Krasula, L., Shahid, M., Temel, D., Brunnström, K., & Razaak, M. (2016) · Electronic Imaging · doi:10.2352/issn.2470-1173.2016.13.iqsp-206 · Cited by 43
Adaptive Hypergraph Convolutional Network for No-Reference 360-degree Image Quality Assessment
Fu, J., Hou, C., Zhou, W., Xu, J., & Chen, Z. (2022) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3503161.3548337 · Cited by 59
“, PSNR, SSIM [5], MS-SSIM [34], FSIM [35], and DeepQA [36]; three learning-based NR 2DIQA metrics, i.”
Blind light field image quality assessment by analyzing angular-spatial characteristics
Cui, Y., Yu, M., Jiang, Z., Peng, Z., & Chen, F. (2021) · Digital Signal Processing · doi:10.1016/j.dsp.2021.103138 · Cited by 43
DEAR-GAN: Degradation-Aware Face Restoration With GAN Prior
Hu, Y., Wang, Y., & Zhang, J. (2023) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2023.3244786 · Cited by 52
ColorVideoVDP: A visual difference predictor for image, video and display distortions
Mantiuk, R., Hanji, P., Ashraf, M., Asano, Y., & Chapiro, A. (2024) · ACM Transactions on Graphics · doi:10.1145/3658144 · Cited by 88
“Achromatic artifacts stemming from graphics pipelines, such as visible blur or contrast loss, can be modelled by existing luminance-only metrics such as SSIM [Wang et al., 2003] or FovVideoVDP [Man-tiuk et al., 2021], but color ones, such as chroma subsampling, cannot.”
BackgroundMethodRSHazeDiff: A Unified Fourier-Aware Diffusion Model for Remote Sensing Image Dehazing
Xiong, J., Yan, X., Wang, Y., Zhao, W., Zhang, X.-P., & Wei, M. (2024) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2024.3487972 · Cited by 64
“M represents the number of scales and we set it to five according to [55].”
BackgroundQuality prediction of asymmetrically distorted stereoscopic images from single views
Wang, J., Zeng, K., & Wang, Z. (2014) · IEEE International Conference on Multimedia and Expo (ICME) · doi:10.1109/icme.2014.6890303 · Cited by 44
Self-citationLearning Receptive Fields and Quality Lookups for Blind Quality Assessment of Stereoscopic Images
Shao, F., Lin, W., Wang, S., Jiang, G., Yu, M., & Dai, Q. (2015) · IEEE Transactions on Cybernetics · doi:10.1109/tcyb.2015.2414479 · Cited by 42
Free-Energy Principle Inspired Video Quality Metric and Its Use in Video Coding
Xu, L., Lin, W., Ma, L., Zhang, Y., Fang, Y., Ngan, K. N., Li, S., & Yan, Y. (2016) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2016.2525004 · Cited by 45
Lv, Y., Yu, M., Jiang, G., Shao, F., Peng, Z., & Chen, F. (2016) · Signal Processing Image Communication · doi:10.1016/j.image.2016.07.003 · Cited by 45
RVSIM: a feature similarity method for full-reference image quality assessment
Yang, G., Li, D.-S., Lu, F., Liao, Y., & Yang, W. (2018) · EURASIP Journal on Image and Video Processing · doi:10.1186/s13640-018-0246-1 · Cited by 43
Fine-Grained Quality Assessment for Compressed Images
Zhang, X., Lin, W., Wang, S., Liu, J., Ma, S., & Gao, W. (2018) · IEEE Transactions on Image Processing · doi:10.1109/tip.2018.2874283 · Cited by 44
“Besides the JPEG default quantization table, the open source and well optimized JPEG codec, libjpeg [42], adopted another 8 quantization tables, and one of them is an optimized quantization table based on MS-SSIM as shown in Fig.1(b).”
InfluentialAn edge preserving high-order PDE for multiframe image super-resolution
Laghrib, A., Hadri, A., & Hakim, A. (2019) · Journal of the Franklin Institute · doi:10.1016/j.jfranklin.2019.02.032 · Cited by 43
Privacy-Net: An Adversarial Approach for Identity-Obfuscated Segmentation of Medical Images
Kim, B. N., Dolz, J., Jodoin, P.-M., & Desrosiers, C. (2021) · IEEE Transactions on Medical Imaging · doi:10.1109/tmi.2021.3065727 · Cited by 43
“We also use the multiscale structural-similarity (MS-SSIM) score to measure image-to-image distance as a proxy of perceived image quality [48].”
MethodRao, R. R. R., Göring, S., List, P., et al. (2020) · International Conference on Quality of Multimedia Experience (QoMEX) · doi:10.1109/qomex48832.2020.9123110 · Cited by 43
“This model was evaluated on its ability to predict pixel-based metrics such as PSNR, PEVQ, SSIM [16] and MS-SSIM [17] and subjective mean opinion scores (MOS).”
MethodRemote Sensing Image Super-Resolution via Saliency-Guided Feedback GANs
Wu, H., Zhang, L., & Ma, J. (2020) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2020.3042515 · Cited by 45
“We adopt four objective metrics to evaluate the quality of super-resolved images, including PSNR, SSIM [24], MS-SSIM [57], and information fidelity criterion (IFC) [58].”
MethodA Subjective and Objective Study of Space-Time Subsampled Video Quality
Lee, D. Y., Paul, S., Bampis, C. G., Ko, H., Kim, J.-H., Jeong, S., Homan, B., & Bovik, A. C. (2021) · IEEE Transactions on Image Processing · doi:10.1109/tip.2021.3137658 · Cited by 47
“The reference models that we evaluated include the image quality assessment (IQA) models such as PSNR, SSIM [30], MSSSIM [31], and VIF [32], computed on each frame yielding predictions that were averaged (pooled) over all frames to obtain overall video quality scores.”
MethodSelf-citationJin, Y., Chen, M., Goodall, T., Patney, A., & Bovik, A. C. (2021) · IEEE Transactions on Image Processing · doi:10.1109/tip.2021.3087322 · Cited by 52
“1) Non-foveated IQA algorithms: We first tested 7 non-foveated IQA algorithms on both databases: PSNR, SSIM [34], MS-SSIM [35], VIF 1 [36], S-RRED [58], Speed-IQA [59] and FSIM [37].”
MethodSelf-citationSurfaceNet: Adversarial SVBRDF Estimation from a Single Image
Vecchio, G., Palazzo, S., & Spampinato, C. (2021) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.01260 · Cited by 43
“4 shows the estimation accuracy when using a supervised loss alone (simply L1 first, and then adding MS-SSIM to obtainLsup), an unsupervised (i.e., adversarial) loss alone (Lunsup), and the overall L loss, both when only synthetic images are used (L “synth”) and when real data are included (L “real”).”
BackgroundInfluentialThe Use of Thermal Cameras for Pedestrian Detection
Altay, F., & Velipasalar, S. (2022) · IEEE Sensors Journal · doi:10.1109/jsen.2022.3172386 · Cited by 54
No-reference image blur assessment using multiscale gradient
Chen, M.-J., & Bovik, A. C. (2009) · International Conference on Quality of Multimedia Experience (QoMEX) · doi:10.1109/qomex.2009.5246973 · Cited by 43
Self-citationEnd-to-end Optimized Image Compression with Attention Mechanism
Zhou, L., Sun, Z., Wu, X., & Wu, J. (2019) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · Cited by 75
PEA265: Perceptual Assessment of Video Compression Artifacts
Lin, L., Yu, S., Zhou, L., Chen, W., Zhao, T., & Wang, Z. (2020) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2020.2980571 · Cited by 47
“The method of P-VQM is also compared with typical video quality metrics including PSNR, SSIM [29], MS-SSIM [47], VIIDEO [48], SpEED-QA [49] and AMB-VQM [33] to show its performance.”
MethodSelf-citationRevise-Net: Exploiting Reverse Attention Mechanism for Salient Object Detection
Hussain, R., Karbhari, Y., Ijaz, M. F., Woźniak, M., Singh, P. K., & Sarkar, R. (2021) · Remote Sensing · doi:10.3390/rs13234941 · Cited by 45
“For x and y, which are the two pixel values corresponding to patches of size N × N that cropped from SM and GT, the SSIM loss [54] can be formulated as in Equation (5).”
MethodIllumination Guided Attentive Wavelet Network for Low-Light Image Enhancement
Xu, J., Yuan, M., Yan, D.-M., & Wu, T. (2022) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2022.3207330 · Cited by 49
“…results and the ground-truth images locally and globally, we minimize the negative multi-scale structure similarity (MS-SSIM) index distance [50] between the enhancement result S E and the normal-light ground-truth image S N , 2) Image Smoothness Loss: To further refine texture details and…”
BackgroundUrbanEvolver: Function-Aware Urban Layout Regeneration
Qin, Y., Zhao, N., Yang, J., Pan, S., Sheng, B., & Lau, R. W. H. (2024) · International Journal of Computer Vision · doi:10.1007/s11263-024-02030-w · Cited by 52
“To obtain the lines, we use a line detector (Pautrat et al., 2021) D L , and compute the differences between the regenerated lines and ground-truth lines in hough space (Zhao et al., 2021), as: where H (.) is the deep hough transformation to transfer the lines from pixel space into parameter space,…”
A multi-metric fusion approach to visual quality assessment
Liu, T.-J., Lin, W., & Kuo, C. ‐. J. (2011) · International Conference on Quality of Multimedia Experience (QoMEX) · doi:10.1109/qomex.2011.6065715 · Cited by 46
“Metric Name MSSSIM [3] SSIM [2] VIF [4] VSNR [5] VIFP [4] NQM [6] UQI [7] PSNR -HVS [8] IFC [9] PSNR”
BackgroundMethod
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.