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 1,801–1,900 of 6,047
System-on-a-chip (SoC)-based hardware acceleration for foreground and background identification
Safaei, A., Wu, Q. M. J., & Yang, Y. (2017) · Journal of the Franklin Institute · doi:10.1016/j.jfranklin.2017.07.037 · Cited by 17
“The multiscale structural similarity index (MS-SSIM) [42] is used to evaluate the precision by considering the structural similarity and comparing it against a ground truth.”
MethodPerceptnet: A Human Visual System Inspired Neural Network For Estimating Perceptual Distance
Hepburn, A., Laparra, V., Malo, J., McConville, R., & Santos-Rodríguez, R. (2019) · International Conference on Information Photonics · doi:10.1109/icip40778.2020.9190691 · Cited by 28
“Models based on structural similarity such as SSIM [13] and its variants MSSSIM and FSIM [5, 7] check the integrity of the statistics of the distorted image.”
MethodInfluentialPerceptual QoE-Optimal Resource Allocation for Adaptive Video Streaming
Eswara, N., Chakraborty, S., Sethuram, H. P., Kuchi, K., Kumar, A., & Channappayya, S. S. (2019) · IEEE Transactions on Broadcasting · doi:10.1109/tbc.2019.2954064 · Cited by 19
“STSQs can be computed using any of the video quality assessment (VQA) methods such as STRRED [47], MS-SSIM [52], VMAF [53] etc.”
MethodExtreme Image Coding via Multiscale Autoencoders with Generative Adversarial Optimization
Huang, C., Liu, H., Chen, T., Shen, Q., & Ma, Z. (2019) · IEEE International Conference on Visual Communications and Image Processing (VCIP) · doi:10.1109/vcip47243.2019.8966059 · Cited by 17
“This is mainly due to the reason that visual sensitive information (i.e., perceptual significance) can not be well preserved using conventional quality optimization criteria, such as peak signal-to-noise ratio (PSNR) and multiscale structural similarity (MS-SSIM) [8], at such extreme compression scenario.”
BackgroundNonlinear Regression Color Correction Method for RGBN Cameras
Han, Z., Jin, W., Li, L., Wang, X., Bai, X., & Wang, H. (2020) · IEEE Access · doi:10.1109/access.2020.2971423 · Cited by 17
“The equation of SSIM is [23] SSIM (f , g) = l(f , g)c(f , g)s(f , g) l(f , g) = 2μf μg + C1 μ(2)f + μ 2 g + C1”
Croci, S., Ozcinar, C., Zerman, E., Knorr, S. B., Cabrera, J., & Smolić, A. (2020) · Quality and User Experience · doi:10.1007/s41233-020-00032-3 · Cited by 20
“Four of the existing metrics that we evaluate were developed for traditional 2D image/video quality assessment: PSNR, SSIM [43], MS-SSIM [44], and VMAF [39].”
MethodTSPR: Deep network-based blind image quality assessment using two-side pseudo reference images
Hu, J., Wang, X., Shao, F., & Jiang, Q. (2020) · Digital Signal Processing · doi:10.1016/j.dsp.2020.102849 · Cited by 19
Ling, S., Li, J., Che, Z., Min, X., Zhai, G., & Callet, P. L. (2020) · IEEE Transactions on Image Processing · doi:10.1109/tip.2020.3037504 · Cited by 19
Evaluation of quality measures for color quantization
Ramella, G. (2020) · Multimedia tools and applications · doi:10.1007/s11042-021-11385-y · Cited by 19
“If M is the total number of windows is possible to compute the Mean Structural SIMilarity index (MSSIM), given as [64]:”
MethodFACER: A Universal Framework for Detecting Anomalous Operation of Deep Neural Networks
Schorn, C., & Gauerhof, L. (2020) · IEEE International Conference on Intelligent Transportation Systems (ITSC) · doi:10.1109/itsc45102.2020.9294226 · Cited by 22
“The severity levels 1, 2, . . . , 5 correspond to an average reduction of the multiscale structural similarity (MS-SSIM) [33] to 0.9, 0.8, . . . , 0.5, respectively, where MS-SSIM is measured between a set of original images and their noisy variants.”
BackgroundInfluentialAn optimized CNN-based quality assessment model for screen content image
Jiang, X., Shen, L., Feng, G., Yu, L., & An, P. (2021) · Signal Processing Image Communication · doi:10.1016/j.image.2021.116181 · Cited by 18
Qualitative Rating of Lossy Compression for Aerial Imagery by Neutrosophic WASPAS Method
Baušys, R., & Kazakevičiūtė-Januškevičienė, G. (2021) · Symmetry · doi:10.3390/sym13020273 · Cited by 17
Comprehensive evaluation of image enhancement for unsupervised image description and matching
Lecca, M., Torresani, A., & Remondino, F. (2020) · IET Image Processing · doi:10.1049/iet-ipr.2020.1129 · Cited by 18
“The structural differences are measured by SSIM [38] and MS-SSIM [39] and they are computed locally and globally, while the visual differences are based on features extracted through a VGG network [40].”
MethodProgressive face super-resolution with cascaded recurrent convolutional network
Liu, S., Xiong, C., Shi, X., et al. (2021) · Neurocomputing · doi:10.1016/j.neucom.2021.03.124 · Cited by 17
Ayyoubzadeh, S. M., & Royat, A. (2021) · 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw53098.2021.00049 · Cited by 14
“To improve, the correlation of SSIM with human perception from image quality, some other variants of that such as RFSIM [41] , SRSIM[39] , FSIM[42] , VSI[40] , MS-SSIM[37] and GSM [22] are introduced.”
MethodDeep learning-based image reconstruction for few-view computed tomography
Yim, D., Lee, S., Nam, K., Lee, D., Kim, D. K., & Kim, J.-S. (2021) · Nuclear Instruments and Methods in Physics Research Section A Accelerators Spectrometers Detectors and Associated Equipment · doi:10.1016/j.nima.2021.165594 · Cited by 19
Handwritten Text Generation via Disentangled Representations
Liu, X., Meng, G., Xiang, S., et al. (2021) · IEEE Signal Processing Letters · doi:10.1109/lsp.2021.3109541 · Cited by 17
Hierarchical Edge Refinement Network for Saliency Detection
Song, D., Dong, Y., & Li, X. (2021) · IEEE Transactions on Image Processing · doi:10.1109/tip.2021.3106798 · Cited by 18
HDR Image Reconstruction Using Segmented Image Learning
Lee, B. D., & Sunwoo, M. H. (2021) · IEEE Access · doi:10.1109/access.2021.3119586 · Cited by 16
“This section compares the results from the proposed and the existing methods using the peak signal noise to rate (PSNR) [47], structural similarity index measure (SSIM) [48], and multiscale SSIM (MS-SSIM) [49] metrics, which are the most common evaluation indicators.”
MethodDecompose kV projection using neural network for improved motion tracking in paraspinal SBRT
He, X., Cai, W., Li, F., Fan, Q., Zhang, P., Cuaron, J. J., Cerviño, L., Li, X., & Li, T. (2021) · Medical Physics · doi:10.1002/mp.15295 · Cited by 17
Variable-Rate Deep Image Compression With Vision Transformers
Li, B., Liang, J., & Han, J. (2022) · IEEE Access · doi:10.1109/access.2022.3173256 · Cited by 25
“MS-SSIM is a complementary metric to evaluate the structural similarity between two images [41].”
BackgroundMethodInfluentialFull-Reference Quality Estimation for Images With Different Spatial Resolutions
Demirtas, A. M., Reibman, A., & Jafarkhani, H. (2014) · IEEE Transactions on Image Processing · doi:10.1109/tip.2014.2310991 · Cited by 20
“Compared state-of-the-art FR QE methods are Structural Similarity Index (SSIM) [8], VIF [9], Visual Signal-to-Noise Ratio (VSNR) [10], PSNR, MultiscaleSSIM (MSSIM) [11], Information Content Weighted SSIM”
MethodNo-Reference Image Quality Assessment for Image Auto-Denoising
Kong, X., & Yang, Q. (2017) · International Journal of Computer Vision · doi:10.1007/s11263-017-1054-2 · Cited by 17
“These metrics are developed based on structure similarity measurement (Wang et al. 2004, 2003; Wang and Li 2011), analysis with singular vector decomposition (Shnayderman et al. 2006; Narwaria and Lin 2012), and image gradient (Zhu and Wang 2012; Chen et al. 2006).”
MethodA nonlinear fourth-order PDE for multi-frame image super-resolution enhancement
Laghrib, A., Chakib, A., Hadri, A., & Hakim, A. (2019) · Discrete and Continuous Dynamical Systems - B · doi:10.3934/dcdsb.2019188 · Cited by 22
A convolutional neural network based method for low-illumination image enhancement
Huang, H., Tao, H., & Wang, H. (2019) · doi:10.1145/3357254.3357255 · Cited by 20
“In this paper, we introduce the PSNR loss and combine it with MS-SSIM loss into the final mix loss.”
MethodInfluentialPerceptual Texture Similarity Estimation: An Evaluation of Computational Features
Dong, X., Dong, J., & Chantler, M. J. (2020) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2020.2964533 · Cited by 18
Learning for Video Compression with Hierarchical Quality and Recurrent Enhancement
Yang, R., Mentzer, F., Gool, L. V., & Timofte, R. (2020) · Repository for Publications and Research Data (ETH Zurich) · doi:10.3929/ethz-b-000454769 · Cited by 18
A Survey on the Progression and Performance of Generative Adversarial Networks
Ghosh, B., Dutta, I. K., Totaro, M. W., & Bayoumi, M. (2020) · International Conference on Computing, Communication and Networking Technologies (ICCCNT) · doi:10.1109/icccnt49239.2020.9225510 · Cited by 21
“Multi-scale structural similarity for image quality (MS-SSIM) [90] This evaluates the structural similarity on by comparing the luminance, contrast and the structure at multiple scales to assess the image quality.”
BackgroundPerceptual Quality of Light Field Images and Impact of Visualization Techniques
Paudyal, P., Battisti, F., Callet, P. L., Gutiérrez, J., & Carli, M. (2020) · IEEE Transactions on Broadcasting · doi:10.1109/tbc.2020.3034445 · Cited by 20
“The considered metrics are UQI [69], SSIM [70], and MS-SSIM [71].”
MethodCombined frequency and spatial domain-based patch propagation for image completion
Florinabel, D. J., Juliet, S. E., & Sadasivam, V. (2011) · Computers & Graphics · doi:10.1016/j.cag.2011.10.002 · Cited by 17
Towards Metamerism via Foveated Style TransferPreprint
Deza, A., Jonnalagadda, A., & Eckstein, M. P. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1705.10041 · Cited by 26
Privacy Partitioning: Protecting User Data During the Deep Learning Inference PhasePreprint
Chi, J., Owusu, E., Yin, X., Yu, T., Chan, W., Tague, P., & Tian, Y. (2018) · arXiv.org · doi:10.48550/arxiv.1812.02863 · Cited by 29
A Holistic Survey of Wireless Multipath Video StreamingPreprint
Afzal, S., Testoni, V., Rothenberg, C. E., Kolan, P., & Bouazizi, I. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1906.06184 · Cited by 18
“We have added some additional video quality metrics, such as Peak Signal-to-Noise Ratio (PSNR), Video Quality Metric (VQM), Structural SIMilarity (SSIM) [198], Multi-Scale Structural SIMilarity (MS-SSIM) [199], and Subjective Assessment Methodology for VIdeo Quality (SAMVIQ) [200].”
MethodDiving Deeper into Underwater Image Enhancement: A SurveyPreprint
Anwar, S., & Li, C. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1907.07863 · Cited by 17
“The channels are then subjected to different size of sliding windows to compute the SSIM loss between the turbid and generated images.”
BackgroundInfluentialAdaptive-CS-Net: FastMRI with Adaptive IntelligencePreprint
Pezzotti, N., de Weerdt, E., Yousefi, S., et al. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1912.12259 · Cited by 17
“Note that the loss is applied only on the center slice and the network is free to learn to reconstruct the neighboring slices only if it helps to better reconstruct xc. MSSIM encourages the network to learn and preserve structural and contrast information, while L1 enforces sensitivity to uniform biases to preserve lu…”
MethodInfluentialAdaptive Structure and Texture Similarity Metric for Image Quality Assessment and Optimization
Ding, K.-Y., Zhong, R., Wang, Z.-H., Yu, Y., & Fang, Y.-M. (2024) · IEEE transactions on multimedia · doi:10.1109/tmm.2023.3333208 · Cited by 17
“3) Baselines: We compare A-DISTS against nine knowledge -drivenIQAmodels,includingPSNR,SSIM[1],MS-SSIM[28], VIF[31],MAD[26],FSIM c [29], GMSD [49], VSI [50], NLPD [27], and five mainstream data-driven IQA models PieAPP [12], LPIPS [11], DISTS [13], DSD [51] and DeepWSD [16].”
MethodSelf-citationGenerative AI Empowered Semantic Feature Multiple Access (SFMA) Over Wireless Networks
Wang, J., Yang, Y., Yang, Z., Huang, C., Chen, M., Zhang, Z., & Shikh‐Bahaei, M. (2025) · IEEE Transactions on Cognitive Communications and Networking · doi:10.1109/tccn.2025.3529692 · Cited by 21
“We utilize the Multi-Scale Structural Similarity Index (MS-SSIM) metric [30] to evaluate structural similarity and the Learned Perceptual Image Patch Similarity (LPIPS) metric [31] to measure the perceptual similarities.”
Debiased Mapping for Full-Reference Image Quality Assessment
Chen, B.-L., Zhu, H.-W., Zhu, L.-Y., Wang, S.-S., Pan, J., & Wang, S.-Q. (2025) · IEEE transactions on multimedia · doi:10.1109/tmm.2025.3535280 · Cited by 18
“Method PSNR SSIM [7] MS-SSIM [17] VIF [21] MAD [39] IW-SSIM [18] IGM [40]”
InfluentialBeyond Privacy: Generating Privacy-Preserving Faces Supporting Robust Image Authentication
Wang, T., Wen, W., Xiao, X., Hua, Z., Zhang, Y., & Fang, Y. (2025) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2025.3541859 · Cited by 26
“SSIM [54], LPIPS [50], and FID [55] are used to calculate the image similarity.”
Unconditional latent diffusion models memorize patient imaging data
Dar, S. U. H., Seyfarth, M., Ayx, I., Papavassiliu, T., Schoenberg, S. O., Siepmann, R. M., Laqua, F. C., Kahmann, J., Frey, N., Baeßler, B., et al. (2025) · Nature Biomedical Engineering · doi:10.1038/s41551-025-01468-8 · Cited by 53
Using Structural degradation and Parallax for reduced-reference quality assessment of 3D images
Xu, Q., Zhai, G., Liu, M., et al. (2014) · IEEE International Symposium on Broadband Multimedia Systems and Broadcasting (BMSB) · doi:10.1109/bmsb.2014.6873484 · Cited by 16
EGB: Image Quality Assessment based on Ensemble of Gradient Boosting
Hammou, D., Fezza, S. A., & Hamidouche, W. (2021) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw53098.2021.00066 · Cited by 22
“Main Score ↑ SROCC ↑ PLCC ↑ PSNR 0.5464 0.2547 0.2916 NQM [7] 0.7621 0.3457 0.4163 UQI [38] 1.0334 0.4858 0.5475 SSIM [41] 0.7383 0.3399 0.3984 MS-SSIM [42] 1.0496 0.4863 0.5632 IFC [34] 1.2703 0.5936 0.6766 VIF [33] 0.9570 0.4334 0.5235 VSNR [5] 0.6962 0.3212 0.3750 RFSIM [46] 0.5700 0.2655 0.3044 GSM [24] 0.8869 0.4…”
MethodInfluentialFashion-Guided Adversarial Attack on Person Segmentation
Treu, M., Le, T.-N., Nguyen, H. H., Yamagishi, J., & Echizen, I. (2021) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw53098.2021.00105 · Cited by 18
“where Lsim is computed as the multi-scale structural similarity index measure (MS-SSIM) between X̃ and X [43], and Ltv stands for the total variation loss, which is helpful for reducing noise in the adversarial texture [29].”
Niu, B., Qu, X., Guan, X., et al. (2021) · Optics Express · doi:10.1364/oe.434950 · Cited by 16
Jaubert, O., Steeden, J. A., Montalt‐Tordera, J., Arridge, S., Kowalik, G. T., & Muthurangu, V. (2021) · Magnetic Resonance Imaging · doi:10.1016/j.mri.2021.08.005 · Cited by 16
DMDN: Degradation model-based deep network for multi-focus image fusion
Xiao, Y., Guo, Z., Veelaert, P., et al. (2021) · Signal Processing Image Communication · doi:10.1016/j.image.2021.116554 · Cited by 16
Spatial Contrastive Learning for Anomaly Detection and Localization
Kim, D.-H., Jeong, D., Kim, H., Chong, K., Kim, S., & Cho, H. (2022) · IEEE Access · doi:10.1109/access.2022.3149130 · Cited by 19
“The anomaly map of the PAE is calculated based on the Structural Similarity Index (SSIM) [41] to get pixel-size features.”
MethodAbazari, M. A., Soltani, M., Kashkooli, F. M., & Raahemifar, K. (2022) · Cancers · doi:10.3390/cancers14112786 · Cited by 19
“In our study, an SSIM index of 0 indicates no similarity, and 1 indicates total positive similarity [42]: Figure 3.”
BackgroundInfluentialStructure-Preserving Motion Estimation for Learned Video Compression
Gao, H., Cui, J., Ye, M., et al. (2022) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3503161.3548156 · Cited by 16
“BDBR [8] value is also calculated, which represents the average bit delta using the same PSNR or MS-SSIM.”
MethodInfluentialHistogram-Based Transformation Function Estimation for Low-Light Image Enhancement
Park, J., Vien, A. G., Kim, J. H., & Lee, C. (2022) · Proceedings - International Conference on Image Processing · doi:10.1109/icip46576.2022.9897778 · Cited by 20
“Table 1 shows the quantitative comparison of the low-light image enhancement performance using five image quality metrics: PSNR, MS-SSIM [18], IW-SSIM [19], FSIM [20], and LPIPS [21].”
MethodSingle Infrared Image Stripe Removal via Residual Attention Network
Ding, D., Li, Y., Zhao, P., et al. (2022) · Sensors · doi:10.3390/s22228734 · Cited by 16
“Direction Structure Similarity Algorithm As a full-reference image similarity metric, the structural similarity algorithm (SSIM) estimates from three different factors: brightness, contrast and structure [25,26].”
MethodDeep-Learning Architecture-Based Approach for 2-D-Simulation of Microwave Plasma Interaction
Desai, M., Ghosh, P., Kumar, A., & Chaudhury, B. (2022) · IEEE Transactions on Microwave Theory and Techniques · doi:10.1109/tmtt.2022.3217138 · Cited by 21
“The results obtained from the network have been evaluated using various metrics such as SSIM and MSE.”
InfluentialLu, X., Quan, W., Gao, S., et al. (2022) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2022.3211558 · Cited by 16
“In addition, to further improve the resolution of generated insulators, in addition to original MSE loss in GANomaly, both feature loss and MS-SSIM [35] loss are together introduced into two-stage training of RIIR-Net. which plays a role in measuring the difference between the original images and…”
MethodReflection Removal With NIR and RGB Image Feature Fusion
Hong, Y., Lyu, Y., Li, S., Cao, G., & Shi, B. (2022) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2022.3217446 · Cited by 18
“For error metrics, we utilize the peak signal-to-noise ratio (PSNR) [45], the structural similarity index (SSIM) [35], the normalized cross correlation (NCC) [46], and the least mean square error (LMSE) [47].”
MethodDeep Quality Assessment of Compressed Videos: A Subjective and Objective Study
Lin, L., Wang, Z., He, J., Chen, W., Xu, Y., & Zhao, T. (2022) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2022.3227039 · Cited by 17
“The method of STFEE is also compared with typical video quality metrics including PSNR, SSIM, MS-SSIM [36], 3D-PSD [37], SpEED-VQA [7], NIQE [38], VIIDEO [23], TLVQM [39], VSFA [12], MDTVSFA [40], GSTVQA [13] and Shen’s algorithm [41] to show its performance.”
MethodSelf-citationMASIC: Deep Mask Stereo Image Compression
Deng, X., Deng, Y., Yang, R., Yang, W., Timofte, R., & Xu, M. (2023) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2023.3253702 · Cited by 23
Li, Q., & Kong, Y. (2023) · Remote Sensing · doi:10.3390/rs15082153 · Cited by 16
“…achieved considerable According to the dependence of the image to be evaluated on the information related to the original image, the objective image quality evaluation methods can be roughly divided into three categories: the complete reference, partial reference, and no reference. success [22–26].”
BackgroundInv3D: a high-resolution 3D invoice dataset for template-guided single-image document unwarping
Hertlein, F., Naumann, A., & Philipp, P. (2023) · International Journal on Document Analysis and Recognition (IJDAR) · doi:10.1007/s10032-023-00434-x · Cited by 19
ST360IQ: No-Reference Omnidirectional Image Quality Assessment With Spherical Vision Transformers
Tofighi, N. J., Elfkir, M. H., İmamoğlu, N., Ozcinar, C., Erdem, E., & Erdem, A. (2023) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49357.2023.10096750 · Cited by 22
Underwater Image Enhancement with Phase Transfer and Attention
Khan, M. R., Kulkarni, A., Phutke, S. S., & Murala, S. (2023) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn54540.2023.10191620 · Cited by 17
“Unlike SSIM, the multi-scale SSIM (MS-SSIM) loss [40] considers different resolutions of input providing more flexibility.”
MethodCA-Wav2Lip: Coordinate Attention-based Speech To Lip Synthesis In The Wild
Wang, K.-C., Zhang, J., Huang, J., Li, Q. L., Sun, M.-T., Sakai, K., & Ku, W. (2023) · IEEE International Conference on Smart Computing · doi:10.1109/smartcomp58114.2023.00018 · Cited by 18
“For visual quality assessment, SSIM [40] and MSSSIM [41] are evaluated.”
MethodComplexity-guided Slimmable Decoder for Efficient Deep Video Compression
Hu, Z., & Xu, D. (2023) · 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/cvpr52729.2023.01380 · Cited by 16
“Considering that we need three metrics ( i.e ., MACs, bpp, PSNR/MS-SSIM) to evaluate our method, we use the BD-PSNR [5] and BD-MSSSIM (dB) to evaluate the average PSNR/MS-SSIM(dB) improvement over two rate-distortion curves.”
MethodAdaptive Global Decay Process for Event Cameras
Nunes, U. M., Benosman, R. B., & Ieng, S.-H. (2023) · 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/cvpr52729.2023.00942 · Cited by 19
“We then measure the corresponding image similarity with the real log-frames gradient magnitude, in terms of Structural Similarity Index (SSIM) [40] and Multi-Scale SSIM (MSSIM) [41], to demonstrate that the proposed method globally adapts to event stream dynamics by yielding sharp edge-like images.”
MethodStrawberry Defect Identification Using Deep Learning Infrared–Visible Image Fusion
Lu, Y., Gong, M., Li, J., & Ma, J. (2023) · Agronomy · doi:10.3390/agronomy13092217 · Cited by 17
Zhao, L., Song, D., Chen, W., & Kang, Q. (2023) · Computer-Aided Civil and Infrastructure Engineering · doi:10.1111/mice.13116 · Cited by 17
Pavarut, S., Preedanan, W., Kumazawa, I., Suzuki, K., Kobayashi, M., Tanaka, H., Ishioka, J., Matsuoka, Y., & Fuji, Y. (2023) · IEEE Access · doi:10.1109/access.2023.3345648 · Cited by 20
“For the quantitative analysis, we employ image quality assessment methods such as Mean Squared Error (MSE), Peak Signal-to-Noise Ratio (PSNR), and Structural Similarity Index (SSIM) [32].”
MethodInnovating Real Fisheye Image Correction with Dual Diffusion Architecture
Yang, S., Lin, C., Liao, K., & Zhao, Y. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.01166 · Cited by 20
Lin, L., Bai, S., Qu, M., Wei, X., Wang, L., Wu, F., Liu, B., Zhou, W., & Kuruoğlu, E. E. (2024) · IEEE Transactions on Broadcasting · doi:10.1109/tbc.2024.3353579 · Cited by 19
“The traditional 2D IQA methods are usually based on various handcrafted features and metrics, such as structural similarity (SSIM) [6], feature similarity [7], multi-scale SSIM [8], visual information fidelity [9], information content weighting SSIM [10], [11], etc.”
MethodEnhancing RAW-to-sRGB with Decoupled Style Structure in Fourier Domain
He, X., Hu, T., Wang, G., Wang, Z., Wang, R., Zhang, Q., Yan, K., Chen, Z., Li, R., Xie, C., et al. (2024) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v38i3.27985 · Cited by 36
Self-citationYu, F., Xie, Y., Wu, L., Wen, Y., Wang, G., Ren, S., Chen, X., Mao, J., & Li, W. (2024) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv57701.2024.00072 · Cited by 29
“For image similarity, we use Multi-Scale Structural Similarity ( MS-SSIM ) [32] and Local Distortion ( LD ) [36].”
MethodNavigating Uncertainty: Semantic-Powered Image Enhancement and Fusion
Yao, J., Zhao, Y., Kong, S. G., & Zhang, X. (2024) · IEEE Signal Processing Letters · doi:10.1109/lsp.2024.3389463 · Cited by 17
“We comprehensively assess them using metrics of entropy (EN), average gradient (AG), visual information fidelity (VIF) [39] and MS-SSIM [40] from the following aspects: information content,”
MethodLSSVC: A Learned Spatially Scalable Video Coding Scheme
Bian, Y., Sheng, X., Li, L., et al. (2024) · IEEE Transactions on Image Processing · doi:10.1109/tip.2024.3395025 · Cited by 16
“We also train the EL model for four different bitrates and different distortion like MS-SSIM [43].”
MethodMS-PCQE: Efficient NR-PCQE via Multi-Scale Interaction Module in Immersive Communications
Chai, X., & Shao, F. (2024) · IEEE Transactions on Consumer Electronics · doi:10.1109/tce.2024.3423830 · Cited by 19
Exploring Generative AI for Sim2Real in Driving Data Synthesis
Zhao, H., Wang, Y., Bashford‐Rogers, T., Donzella, V., & Debattista, K. (2024) · IEEE Intelligent Vehicles Symposium · doi:10.1109/iv55156.2024.10588493 · Cited by 30
“1) Image quality: two sets of image quality metrics are applied, which include the reference-based metrics (i.e. PSNR [44], SSIM [45], LPIPS [46], CW-SSIM [47], FSIM [48]) and non-reference-based image quality metrics (i.e. BRISQUE [49], NIQE [50], FID [51]).”
Perception-Distortion Balanced Super-Resolution: A Multi-Objective Optimization Perspective
Sun, L., Liang, J., Liu, S., Yong, H., & Zhang, L. (2024) · IEEE Transactions on Image Processing · doi:10.1109/tip.2024.3434426 · Cited by 29
“The former adopts the widely used image quality assessment index, i . e ., SSIM [14, 16], as the optimization objective, while the latter Fig.”
MethodChen, B., Yu, Y., Zhang, L., Li, Z., Zhu, C., Yu, C., Song, C., Liu, N., & Liu, Z. (2024) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2024.3492505 · Cited by 21
Aziz, M., Rehman, U., Danish, M. U., & Grolinger, K. (2025) · IEEE Transactions on Human-Machine Systems · doi:10.1109/thms.2025.3527397 · Cited by 35
“MS-SSIMisparticularlyeffectiveincontextswhereimagesmaybe viewedatmultiplescalesorwhenmoresensitivitytochanges instructuralinformationisrequired[12].”
Full Reference Printed Image Quality: Measurement Framework and Statistical Evaluation
Eerola, T., Lensu, L. T., Kalviainen, H. A., Kämäräinen, J., Leisti, T., Nyman, G., Halonen, R., & Oittinen, P. T. (2010) · Journal of Imaging Science and Technology · doi:10.2352/j.imagingsci.technol.2010.54.1.010201 · Cited by 21
Reducing complexity in polygonal meshes with view-based saliency
Castelló, P., Chover, M., Sbert, M., & Feixas, M. (2014) · Computer Aided Geometric Design · doi:10.1016/j.cagd.2014.05.001 · Cited by 15
No-reference quality assessment of H.264/AVC encoded video based on natural scene features
Zhu, K., Asari, V. K., & Saupe, D. (2013) · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · doi:10.1117/12.2015594 · Cited by 18
Sparse Structural Similarity for Objective Image Quality Assessment
Zhang, X., Wang, S., Gu, K., Jiang, T., Ma, S., & Gao, W. (2015) · Conference proceedings/Conference proceedings - IEEE International Conference on Systems, Man, and Cybernetics · doi:10.1109/smc.2015.276 · Cited by 16
“Thus, more accurate criteria considering HVS characteristics have been studied, including visual signal-tonoise ratio (VSNR) [4], internal generative mechanism (IGM) [5], structure similarity (SSIM) [6] and its variants MS-SSIM [7], IW-SSIM [8].”
BackgroundMethodOn Hypothesis Testing for Comparing Image Quality Assessment Metrics [Tips & Tricks]
Zhu, R., Zhou, F., Yang, W.-M., & Xue, J.-H. (2018) · IEEE Signal Processing Magazine · doi:10.1109/msp.2018.2829209 · Cited by 13
Impacts of Retina-Related Zones on Quality Perception of Omnidirectional Image
Tran, H. T. T., Nguyen, D. V., Ngoc, N. P., et al. (2019) · IEEE Access · doi:10.1109/access.2019.2953983 · Cited by 16
“MS-SSIM [18] No No Multi-scale SSIM, Calculated based on similar measures computed at different resolutions (or multi-scales) of a viewport”
MethodGuan, Y., Sang, X., Xing, S., et al. (2019) · IEEE Access · doi:10.1109/access.2019.2956102 · Cited by 16
No-Reference Video Quality Assessment Based on Visual Memory Modeling
Banitalebi-Dehkordi, M., Ebrahimi-Moghadam, A., Khademi, M., & Hadizadeh, H. (2019) · IEEE Transactions on Broadcasting · doi:10.1109/tbc.2019.2957670 · Cited by 17
Soft Video Multicasting Using Adaptive Compressed Sensing
Hadizadeh, H., & Bajić, I. V. (2020) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2020.2975420 · Cited by 15
Ultra Low Bitrate Learned Image Compression by Selective Detail Decoding
Akutsu, H., Suzuki, A., Zhong, Z., et al. (2020) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw50498.2020.00067 · Cited by 16
“In past research, PSNR (MSE), MS-SSIM [13], etc.”
Building Super-Resolution Image Generator for OCR Accuracy Improvement
Peng, X., & Wang, C. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-57058-3_11 · Cited by 16
Xie, Q., Zhao, Q., Xu, Z., et al. (2020) · Science China Information Sciences · doi:10.1007/s11432-020-2880-3 · Cited by 16
“Four quantitative picture quality indices (PQI) are employed for performance evaluation, including peak signal-to-noise ratio (PSNR), structure similarity (SSIM [40]), feature similarity (FSIM [41]), and multiscale structural similarity (MS-SSIM [42]).”
MethodFang, Y., Sui, X., Wang, J., Yan, J., Lei, J., & Callet, P. L. (2020) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2020.3035679 · Cited by 16
Hardware and software video encoding comparison
Safin, R., Garipova, E., Lavrenov, R., et al. (2020) · Annual Conference of the Society of Instrument and Control Engineers of Japan (SICE) · doi:10.23919/sice48898.2020.9240439 · Cited by 16
“3 [21]:”
Reversible Data Hiding in JPEG Images Under Multi-Distortion Metric
Chen, K., Zhou, H., Hou, D., et al. (2020) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2020.3044186 · Cited by 16
“The peak signal-to-noise ratio (PSNR) and structural similarity (SSIM) [41], calculated between the original JPEG image and the marked JPEG image, are used as measures to evaluate the visual quality of the marked JPEG image.”
MethodZhang, X., & Yan, H. (2021) · IET Image Processing · doi:10.1049/ipr2.12137 · Cited by 15
“normalized mutual information (NMI) index [54], gradient-based (QG ) index [55], feature similarity (FSIM) index [56], multi-scale structural similarity (MS-SSIM) index [57] and visual information fidelity for fusion (VIFF) [58] are used to evaluate the quality of fusion and denoising results.”
MethodNo-reference light field image quality assessment based on depth, structural and angular information
Xiang, J., Jiang, G., Yu, M., Bai, Y., & Zhu, Z. (2021) · Signal Processing · doi:10.1016/j.sigpro.2021.108063 · Cited by 16
Spatiotemporal Reflectance Fusion Using a Generative Adversarial Network
Shang, C., Li, X., Yin, Z., Li, X., Wang, L., Zhang, Y., Du, Y., & Ling, F. (2021) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2021.3065418 · Cited by 16
“…the ability of MSE-based loss in the generative network to capture realistic perceptual details, such as high-level spatial textures, is very limited, because these MSE-oriented approaches fundamentally disagree with the evaluation of realistic spatial shapes, as discussed in Section II-B [48].”
BackgroundSkarsoulis, K., Kakkava, E., & Psaltis, D. (2021) · Optics Communications · doi:10.1016/j.optcom.2021.126968 · Cited by 16
Decoupled Representation Learning for Character Glyph Synthesis
Liu, X., Meng, G., Chang, J., Hu, R., Xiang, S., & Pan, C. (2021) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2021.3072449 · Cited by 20
No-reference screen content video quality assessment
Li, T., Min, X., Zhu, W., Xu, Y., & Zhang, W. (2021) · Displays · doi:10.1016/j.displa.2021.102030 · Cited by 17
Identification of weld geometry from ultrasound scan data using deep learning
Provençal, E., & Laperrière, L. (2021) · Procedia CIRP · doi:10.1016/j.procir.2021.11.021 · Cited by 20
Unsupervised Low-Light Image Enhancement by Extracting Structural Similarity and Color Consistency
Shi, Y., Wang, B., Wu, X., & Zhu, M. Q. (2022) · IEEE Signal Processing Letters · doi:10.1109/lsp.2022.3163686 · Cited by 19
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.