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 2,801–2,900 of 6,047
Image encryption with 6D hyperchaotic system and vision transformer autoencoder
Maity, A., Chen, Y.-L., Alexan, W., Yee, P. L., & Dhara, B. C. (2026) · Scientific Reports · doi:10.1038/s41598-025-34378-5 · Cited by 17
StableCodec: Taming One-Step Diffusion for Extreme Image Compression
Zhang, T., Luo, X., Li, L., & Liu, D. (2025) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.01614 · Cited by 56
“Meanwhile, we use PSNR and MS-SSIM [64] to measure the reconstruction fidelity.”
Alpha trimmed correlation for touchless finger image mosaicing
Rao, S. P., Rajendran, R., Agaian, S. С., & Mulawka, M. (2016) · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · doi:10.1117/12.2224392 · Cited by 7
Bridging the Gap Between Imaging Performance and Image Quality Measures
Fry, E. W. S., Triantaphillidou, S., Jacobson, R. E., Jarvis, J. R., & Jenkin, R. (2018) · Electronic Imaging · doi:10.2352/issn.2470-1173.2018.12.iqsp-231 · Cited by 9
Neural Network-Based Estimation of Distortion Sensitivity for Image Quality Prediction
Bosse, S., Becker, S., Fisches, Z. V., Samek, W., & Wiegand, T. (2018) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2018.8451261 · Cited by 7
“FR IQMs classically follow two main strategies [2]: While bottom-up approaches explicitly model the human visual system (HVS) [3, 4, 5], top-down approaches mimic potentially hypothesized abstract properties of the HVS from a signal processing perspective [6, 7, 8, 9].”
BackgroundMethodInfluentialTranslation of Cellular Protein Localization Using Convolutional Networks
Shigene, K., Hiasa, Y., Otake, Y., Soufi, M., Janewanthanakul, S., Nishimura, T., Sato, Y., & Suetsugu, S. (2021) · Frontiers in Cell and Developmental Biology · doi:10.3389/fcell.2021.635231 · Cited by 7
“…α , β , and γ are equal values in each scale, and β 1 = γ 1 = 0.0448, β 2 = γ 2 = 0.2856, β 3 = γ 3 = 0.3001, β 4 = γ 4 = 0.2363, α 5 = β 5 = γ 5 = 01333, which is derived from the Gaussian distribution with the assumption that the medium resolution is suitable for recognition (Wang et al., 2003).”
MethodInfluentialLearning to Denoise Gated Cardiac PET Images Using Convolutional Neural Networks
Gambin, J. R., Tadi, M. J., Teuho, J., Klén, R., Knuuti, J. M., Koskinen, J., Saraste, A., & Lehtonen, E. (2021) · IEEE Access · doi:10.1109/access.2021.3122194 · Cited by 8
Rate-constrained learning-based image compression
Guerin, N. D., da Silva, R. C., de Oliveira, M. C., et al. (2021) · Signal Processing Image Communication · doi:10.1016/j.image.2021.116544 · Cited by 7
Online Meta Adaptation for Variable-Rate Learned Image Compression
Jiang, W., Wang, W., Li, S., & Liu, S. (2022) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw56347.2022.00065 · Cited by 7
Intensity-Sensitive Similarity Indexes for Image Quality Assessment
Li, X., & Armour, W. (2022) · Proceedings - International Conference on Pattern Recognition/Proceedings/International Conference on Pattern Recognition · doi:10.1109/icpr56361.2022.9956093 · Cited by 12
“To evaluate similarity at different resolutions or scales, multi-scale structural similarity (MS-SSIM) [5] has generalised the single-scale SSIM to multi-scale cases.”
MethodHiding images within audio using deep generative model
Paul, S., & Mishra, D. (2022) · Multimedia Tools and Applications · doi:10.1007/s11042-022-13034-4 · Cited by 9
“For two images X and Y , these functions can be visualized by (12) [32].”
BackgroundBreaking Robust Data Hiding in Online Social Networks
Ping, W., Zhu, Z., Luo, G., et al. (2022) · IEEE Signal Processing Letters · doi:10.1109/lsp.2022.3233539 · Cited by 7
“Four metrics, R_error, the PSNR [24], Ms-ssim(or SSIM) [25] and Niqe [26], are introduced to evaluate the ability of the model to break data hiding.”
On the Design of Approximate Sobel Filter
Aoun, A., Masadeh, M., & Tahar, S. (2022) · doi:10.1109/icm56065.2022.10005354 · Cited by 7
“The objective assessment is measured using the multiscale structural similarity (MS-SSIM) index [15] which measures the quality of the input image compared to the reference image.”
MethodBlock Importance Mapping for Video Encoding
Enhorn, J., Hollmann, C., Sjöberg, R., et al. (2022) · IEEE International Conference on Visual Communications and Image Processing (VCIP) · doi:10.1109/vcip56404.2022.10008857 · Cited by 7
Dual-stream coupling network with wavelet transform for cross-resolution person re-identification
Sun, R., Yang, Z., Zhao, Z., et al. (2023) · Journal of Systems Engineering and Electronics · doi:10.23919/jsee.2023.000028 · Cited by 7
Region-Aware Portrait Retouching With Sparse Interactive Guidance
Zeng, H., Huang, J., Li, J., & Xiong, Z. (2023) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2023.3262185 · Cited by 7
“Objective Evaluation Metrics: To evaluate the performance of our automatic retouching branch and compare with state-of-the-art methods, we adopt the following evaluation metrics: PSNR, SSIM [51], MS-SSIM [52] and the CIELAB color difference Δ E ab [2].”
MethodFingerprint image denoising and inpainting using generative adversarial networks
Zhong, W., Mao, L., & Yang, N. (2023) · Evolutionary Intelligence · doi:10.1007/s12065-023-00850-2 · Cited by 7
“The larger the SSIM value, the higher the structural similarity between the two images [36].”
Enhancement of Ambient Mass Spectrometry Imaging Data by Image Restoration
Xiang, Y., Metodiev, M., Wang, M., Cao, B., Bunch, J., & Takáts, Z. (2023) · Metabolites · doi:10.3390/metabo13050669 · Cited by 7
Development of Artificial Intelligence-Based Dual-Energy Subtraction for Chest Radiography
Yamazaki, A., Koshida, A., Tanaka, T., Seki, M., & Ishida, T. (2023) · Applied Sciences · doi:10.3390/app13127220 · Cited by 8
Visually Guided Sound Source Separation With Audio-Visual Predictive Coding
Song, Z., & Zhang, Z. (2023) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2023.3288022 · Cited by 11
“Visualization of visual sound separation results, from Sound-of-Pixels [23], MP-Net [25], AVPC, and AVPC-RCoP. assessment indices, the peak signal-to-noise ratio (PSNR) and multiscale structure similarity (MS-SSIM) [74], to evaluate the predicted mask images’ quality.”
BackgroundBogacki, P., & Dziech, A. (2023) · IEEE Access · doi:10.1109/access.2023.3289162 · Cited by 7
“…of two images being compared σ x , σ y - standard deviations of two images σ xy - cross-covariance between both images C 1 , C 2 - two constants to stabilize the division Multi-Scale Structural Similarity (MS-SSIM) - it is based on multiple SSIM measurements at different image scales [78].”
MethodExploring the Influence of View and Camera Path Selection for Dynamic Mesh Quality Assessment
Yang, K., Yang, Q., Jung, J., et al. (2023) · IEEE International Conference on Multimedia and Expo (ICME) · doi:10.1109/icme55011.2023.00424 · Cited by 7
Gu, Z., Wu, Z., & Dai, N. (2023) · PLoS ONE · doi:10.1371/journal.pone.0291728 · Cited by 7
Sun, W., Li, P., Liang, Y., Feng, Y., & Zhao, L. (2023) · Bioengineering · doi:10.3390/bioengineering10111288 · Cited by 8
“A typical approach involves comparing the distortion metric with an ideal imaging model or a perfect reference image [4].”
BackgroundCalligraphy Font Generation via Explicitly Modeling Location-Aware Glyph Component Deformations
Zhao, M., Qi, X., Hu, Z., Li, L., Zhang, Y., Huang, Z. H., & Yu, X. (2023) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2023.3342690 · Cited by 8
“RMSE [48], SSIM [49], and MS-SSIM [50] are employed to measure whether pixel-level details can be preserved.”
MethodTemplate-guided Illumination Correction for Document Images with Imperfect Geometric Reconstruction
Hertlein, F., & Naumann, A. (2023) · IEEE/CVF International Conference on Computer Vision Workshops (ICCVW) · doi:10.1109/iccvw60793.2023.00097 · Cited by 7
“MS-SSIM Multiscale structural similarity metric [28] is an established image similarity metric that calculates statistical properties at multiple scales and thus, ensures scaleinvariance.”
BackgroundMethodADNet: An Asymmetric Dual-Stream Network for RGB-T Salient Object Detection
Fang, Y., Hou, R., Bei, J., Ren, T., & Wu, G. (2023) · ACM International Conference on Multimedia in Asia (MMAsia) · doi:10.1145/3595916.3626440 · Cited by 8
“…work [12, 37], we compute these three loss functions to measure the gap between the predicted results and the true results: where BCE loss stands for binary cross entropy loss [3], SSIM loss represents structural similarity index measure [36], IoU loss denotes intersection over union loss [20].”
MethodHolographic Reconstruction With All-Acoustic Diffractive Network
Gu, W., Wang, J., Chai, S., Tran, T. N. H. T., Ta, D., & Liu, X. (2024) · IEEE Transactions on Computational Imaging · doi:10.1109/tci.2024.3356867 · Cited by 7
“When we set the weight value according to the previous super-resolution work [37], [38] to α = 0.16 and β = 0.84, the ADNN-AH tends Considering the mean value can be used to reflect the recovery capability in most extents [15], [39], here, α = 0.16, β = 0.84, γ = 1 is finally set in the training…”
BackgroundMethodInfluentialVoxel-Wise Medical Image Generalization for Eliminating Distribution Shift
Li, F., Wang, Y., Beyan, O., Schöneck, M., & Caldeira, L. (2024) · ACM Transactions on Knowledge Discovery from Data · doi:10.1145/3643034 · Cited by 11
“MS-SSIM [ 28 ] extends SSIM by considering image structures across multiple scales, providing a more comprehensive assessment of perceptual image similarity by accounting for details at various levels.”
MethodPerceptual Quality-Oriented Rate Allocation via Distillation from End-to-End Image Compression
Yang, R., Liu, D., Ma, S., Wu, F., & Gao, W. (2024) · ACM Transactions on Multimedia Computing Communications and Applications · doi:10.1145/3650034 · Cited by 7
“Such metrics include Structural SIMilarity (SSIM ) [33], Multi-Scale Structural SIMilarity (MS-SSIM) [34], Learned Perceptual Image Patch Similarity (LPIPS) [43], and so on.”
BackgroundZhou, H., Tao, H., Duan, Q., Hu, Z., & Deng, Y. (2024) · Multimedia Tools and Applications · doi:10.1007/s11042-024-19691-x · Cited by 7
“…, and L ms − ssim as the reconstruction loss, the segmentation loss (we use the binary cross entropy loss in the experiment), the consistency loss, the l 1 loss function, the perceptual loss function [42], the style loss function [43], the multi-scale structural similarity loss [44], respectively.”
Lee, S., Kang, J., Lee, S., Lin, W., & Bovik, A. C. (2024) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2024.3422490 · Cited by 8
“Projection-basedapproachesfortexturedmeshqualityas-sessmentoftenusewell-knownimagequalitymodelslike, SSIM[13],MSSSIM[11],iCID[43],HDR-VDP2[44].”
BackgroundMethodInfluentialSelf-citationEnhancing QoE for Multi-Device Video Delivery: A Novel Dataset and Model Perspective
Yang, H., Lin, T., Zhang, Y., Xu, Y., Chen, Z., & Yan, J. (2024) · IEEE Transactions on Broadcasting · doi:10.1109/tbc.2024.3443544 · Cited by 9
“1) VQA (Video Quality Assessment): Video quality features such as MS-SSIM [34], PSNR [35], VMAF [36], STRRED [37], and NIQE [38].”
Convex Hull Prediction for Adaptive Video Streaming by Recurrent Learning
Paul, S., Norkin, A., & Bovik, A. C. (2024) · IEEE Transactions on Image Processing · doi:10.1109/tip.2024.3455989 · Cited by 15
“2 Using the same process, as described for VMAF, we also computed a second set of convex hulls using Multi-scale structural similarity (MS-SSIM) [49], another perceptual model that is ubiquitously used for quality assessment.”
MethodSelf-citationRes-NeRV: Residual Blocks For A Practical Implicit Neural Video Decoder
Tarchouli, M., Guionnet, T., Rivière, M., et al. (2024) · Proceedings - International Conference on Image Processing · doi:10.1109/icip51287.2024.10647402 · Cited by 7
Perceptual Evaluation of Audio-Visual Synchrony Grounded in Viewers’ Opinion Scores
Goncalves, L., Mathur, P., Lavania, C., Cekic, M., Federico, M., & Han, K. J. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-72986-7_17 · Cited by 14
Adaptive Surveillance Video Compression With Background Hyperprior
Zhao, Y. Q., Tang, S. Y., & Ye, M. (2024) · IEEE Signal Processing Letters · doi:10.1109/lsp.2024.3521663 · Cited by 7
MagShadow: Physical Adversarial Example Attacks via Electromagnetic Injection
Liu, Z., Lin, F., Ba, Z., Lü, L., & Ren, K. (2025) · IEEE Transactions on Dependable and Secure Computing · doi:10.1109/tdsc.2025.3529197 · Cited by 11
“To demonstrate the effectiveness of perturbation calibration, we use the metric of Multi-Scale Structural Similarity Index (MS-SSIM) [63] to measure the performance of our modeling.”
UVC: A Unified Deep Video Compression Framework
Tang, L., Zhang, X., & Zhang, L. (2025) · ACM Transactions on Multimedia Computing Communications and Applications · doi:10.1145/3715144 · Cited by 8
“The L 𝑅𝐷 is: where 𝑑 ( 𝑋 𝑡 + 3 , ˆ 𝑋 + 3 ) means the distortion between 𝑋 𝑡 + 3 and ˆ 𝑋 + 3 , and we use MSE or Multi-scale Structure Similarity (MS-SSIM) [66] loss.”
Peng, Z., Chen, F., Jiang, G., et al. (2015) · Journal of Visual Communication and Image Representation · doi:10.1016/j.jvcir.2015.10.003 · Cited by 7
A new image quality metric based on MIx-Scale transform
Liu, M., Zhai, G., Gu, K., Xu, Q., Yang, X., Sun, X., Chen, W., & Zuo, Y. (2013) · SiPS ... design and implementation - IEEE Workshop on Signal Processing Systems · doi:10.1109/sips.2013.6674517 · Cited by 7
New image quality assessment method using wavelet leader pyramids
Chen, X., Yang, X., Zheng, S., Lin, W., Zhang, R., & Zhai, G. (2011) · Optical Engineering · doi:10.1117/1.3591949 · Cited by 7
Image quality measure using the phase quantization code
Kim, D.-O., & Park, R. (2010) · IEEE Transactions on Consumer Electronics · doi:10.1109/tce.2010.5506023 · Cited by 7
“Box 1142, Seoul 100-611, Korea (e-mail: rhpark@sogang.ac.kr). (MSSSIM) was proposed [ 6 ].”
Objective Quality Metrics in Correlation with Subjective Quality Metrics for Steganography
Wazirali, R., Slehat, S. S., Chaczko, Z., Borowik, G., & Carrión, L. (2015) · Asia-Pacific Conference on Computer Aided System Engineering (APCASE) · doi:10.1109/apcase.2015.49 · Cited by 7
“UQI(c, s) = L(c, s), C(c, s), S(c, s) (8)”
Structured Feature Similarity with Explicit Feature Map
Kobayashi, T. (2016) · 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.2016.136 · Cited by 7
A new SVD-based image quality assessment
Esmaeilpour, M., Mansouri, A., & Mahmoudi-Aznaveh, A. (2013) · Iranian Conference on Machine Vision and Image Processing · doi:10.1109/iranianmvip.2013.6780013 · Cited by 7
“For instance, recently SSIM [1], as one of the well known image quality assessment method, has been utilized in performance optimization of h.264 video compression standard [2-3].”
MethodInfluentialЛукин, В. В., Zemliachenko, A., Abramov, S., et al. (2016) · European Workshop on Visual Information Processing · doi:10.1109/euvip.2016.7764581 · Cited by 7
No reference quality assessment for stereoscopic images by statistical features
Fang, Y., Yan, J., & Wang, J. (2017) · International Conference on Quality of Multimedia Experience (QoMEX) · doi:10.1109/qomex.2017.7965630 · Cited by 8
Image quality assessment using block-based weighted SVD
Azar, F. T., & Parkkinen, J. P. S. (2018) · Signal Image and Video Processing · doi:10.1007/s11760-018-1287-8 · Cited by 7
“The MATLAB codes for IWPSNR, SSIM, MS-SSIM, VIF, G-SSIM, RFSIM, FSIM, VSI and RF are from [1,2,4,5,7–9, 16 ], respectively.”
MethodA New Objective Image Quality Assessment Metric: For Color and Grayscale Images
Uthayakumar, J., Thirumal, V., & Ponnurangam, D. (2018) · 3D Research · doi:10.1007/s13319-018-0180-0 · Cited by 7
Combining Quality Metrics for Improved HDR Image Quality Assessment
Choudhury, A., & Daly, S. J. (2019) · IEEE International Conference on Multimedia Information Processing and Retrieval (MIPR) · doi:10.1109/mipr.2019.00039 · Cited by 7
“Amongst LDR metrics, MS-SSIM calibrated in PQ domain performs the best for database 1 and VIFp in PQ domain works the best for database 2.”
MethodInfluentialGamut-Extension Methods Considering Color Information Restoration
Takeuchi, M., Sakamoto, Y., Yokoyama, R., Sun, H., Matsuo, Y., & Katto, J. (2019) · IEEE Access · doi:10.1109/access.2019.2923400 · Cited by 8
A Food Dish Image Generation Framework Based on Progressive Growing GANs
Wang, S., Gao, H., Zhu, Y., Zhang, W., & Chen, Y. (2019) · Lecture notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering · doi:10.1007/978-3-030-30146-0_22 · Cited by 7
No-reference quality assessment for image-based assessment of economically important tropical woods
Rajagopal, H., Mokhtar, N. B., Izam, T. F. T. M. N., & Khairunizam, W. (2020) · PLoS ONE · doi:10.1371/journal.pone.0233320 · Cited by 9
“Therefore, WNR-IQA displayed a higher performance compared to BRISQUE, deepIQA, DB-CNN, SSIM, MS-SSIM, FSIM, IW-SSIM, and GMSD.”
MethodInfluentialConversion of a Conventional Wheelchair into an Autonomous Personal Transportation Testbed
Sezer, V., Zengin, R. S., Houshyari, H., & Yilmaz, M. C. (2020) · IntechOpen eBooks · doi:10.5772/intechopen.93117 · Cited by 7
No-Reference Nonuniform Distorted Video Quality Assessment Based on Deep Multiple Instance Learning
Qian, L., Pan, T., Zheng, Y., Zhang, J., Li, M., Yu, B., & Wang, B. (2020) · IEEE Multimedia · doi:10.1109/mmul.2020.3034338 · Cited by 8
CNN-Enabled Visibility Enhancement Framework for Vessel Detection under Haze Environment
Lu, Y., Guo, Y., & Liang, M. (2021) · Journal of Advanced Transportation · doi:10.1155/2021/5598390 · Cited by 7
Multi-Metric Fusion Network for Image Quality Assessment
Peng, Y., Xu, J., Luo, Z., Zhou, W., & Chen, Z. (2021) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw53098.2021.00205 · Cited by 7
“MS-SSIM [14], IW-SSIM[13], FSIM [17]) show better correlation with human perception than simple error visibility methods.”
MethodDetecting Deepfake Videos Using Euler Video Magnification
Das, R., Negi, G., & Smeaton, A. F. (2021) · Electronic Imaging · doi:10.2352/issn.2470-1173.2021.4.mwsf-272 · Cited by 18
A no-reference 3D virtual reality image quality assessment algorithm based on saliency statistics
Poreddy, A. K. R., Kara, P. A., Appina, B., & Simon, A. (2021) · Optics and Photonics for Information Processing · doi:10.1117/12.2597327 · Cited by 7
3D-DWT cross-band statistics and features for No-Reference Video Quality Assessment (NR-VQA)
Vishwakarma, A. K., & Bhurchandi, K. M. (2021) · Optik · doi:10.1016/j.ijleo.2021.167774 · Cited by 7
A Measurement for Distortion Induced Saliency Variation in Natural Images
Yang, X., Li, F., & Liu, H. (2021) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2021.3108538 · Cited by 7
Stereoscopic Video Quality Assessment Using Modified Parallax Attention Module
Imani, H., Zaim, S., Islam, M. B., et al. (2021) · Lecture notes in mechanical engineering · doi:10.1007/978-3-030-90421-0_4 · Cited by 7
Semantic Residual Pyramid Network for Image Inpainting
Luo, H., & Zheng, Y. (2022) · Information · doi:10.3390/info13020071 · Cited by 7
“Among them, the data from the MS-SSIM [33] improved more than other evaluation indicators, which verifies the effectiveness of the added multi-scale components.”
MethodResultInfluentialTian, C., Yang, J., Li, P., Zhang, S., & Mi, S. (2022) · Medical Physics · doi:10.1002/mp.15580 · Cited by 7
Haze Level Evaluation Using Dark and Bright Channel Prior Information
Chu, Y., Chen, F., Fu, H. C., & Yu, H. (2022) · Atmosphere · doi:10.3390/atmos13050683 · Cited by 10
An Effective Foveated 360° Image Assessment Based on Graph Convolution Network
Hương, T. T., Ha, D. T., Tran, H. T. T., Viet, N. D., Tien, B. D., Thanh, N. H., Thang, T. C., & Nam, P. N. (2022) · IEEE Access · doi:10.1109/access.2022.3204766 · Cited by 7
CA-BASNet: A Building Extraction Network in High Spatial Resolution Remote Sensing Images
Huang, L., Zhu, J., Qiu, M., et al. (2022) · Sustainability · doi:10.3390/su141811633 · Cited by 7
Saliency-based point cloud quality assessment method using aware features learning
Laazoufi, A., & Hassouni, M. E. (2022) · International Conference on Wireless Networks and Mobile Communications (WINCOM) · doi:10.1109/wincom55661.2022.9966464 · Cited by 8
“[12] 0.3654 0.2789 2.2440 SSIM-MS-IQA [20] 0.3659 0.2592 2.2437 SSIM-IW-IQA [21] 0.4339 0.3285 2.1720 FSIM-IQA [22] 0.3196 0.3019 2.2843 VIF-IQA [23] 0.5243 0.5647 2.0653 3D-to-2D-PCQA [13] 0.6076 0.6020”
InfluentialMandarin Singing Voice Synthesis with Denoising Diffusion Probabilistic Wasserstein GAN
Cho, Y.-P., Tsao, Y., Wang, H., & Liu, Y.-W. (2022) · Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC) · doi:10.23919/apsipaasc55919.2022.9980018 · Cited by 11
Simulation-based test methods with an automotive camera-in-the-loop for automated driving algorithms
Reway, F., Drechsler, M. F., Murthy, R., et al. (2022) · International Conference on Electrical, Computer, Communications and Mechatronics Engineering (ICECCME) · doi:10.1109/iceccme55909.2022.9988437 · Cited by 7
“Structural Similarity Index (SSIM): This metric combines local image structure, luminance, and contrast into a single local quality score [18].”
BackgroundMethodCNN-Based Post-Processing Filter for Video Compression with Multi-Scale Feature Representation
Qi, Z., Jung, C., Liu, Y., & Li, M. (2022) · IEEE International Conference on Visual Communications and Image Processing (VCIP) · doi:10.1109/vcip56404.2022.10008797 · Cited by 8
“We use two metrics for evaluation: Peak signal-to-noise ratio (PSNR) and multiscale structural similarity (MS-SSIM) [26].”
MethodPerceptual impact of the loss function on deep-learning image coding performance
Mohammadi, S., & Ascenso, J. (2022) · Picture Coding Symposium · doi:10.1109/pcs56426.2022.10018061 · Cited by 7
“Short Description MSE Measures pixel-wise squared differences SSIM [14] Measures the degradation in structural information MS-SSIM [15] SSIM extension that supports variations in image resolution and viewing conditions FSIM [16] Exploits phase congruency and gradient information GMSD [17] Measures…”
MethodEFID: Edge-Focused Image Denoising Using a Convolutional Neural Network
Holla, K. S., Park, N., & Lee, B. (2023) · IEEE Access · doi:10.1109/access.2023.3239835 · Cited by 7
Multiple Levels Perceptual Noise Backed Visual Information Fidelity for Picture Quality Assessment
Peng, C., Wu, M., & Liu, K. (2022) · International Symposium on Intelligent Signal Processing and Communication Systems (ISPACS) · doi:10.1109/ispacs57703.2022.10082853 · Cited by 7
“• Multi-SSIM (MS-SSIM) [30], which proposes a method for multi-scale structural similarity that is more flexible to incorporate variations in the observed conditions.”
MethodReGANIE: Rectifying GAN Inversion Errors for Accurate Real Image Editing
Li, B., Ma, T., Zhang, P., Hua, M., Liu, W., He, Q., & Yi, Z. (2023) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v37i1.25210 · Cited by 7
Bombini, A., Bofías, F. G.-A., Ruberto, C., et al. (2023) · RENDICONTI LINCEI · doi:10.1007/s12210-023-01174-0 · Cited by 7
“Following a customary choice in computer vision, all the networks inferring an RGB colour code are trained using the Binary Cross Entropy (BCE) as loss function, while their performance is assessed through several Image Quality measures (Zhao et al. 2015), such as Peak Signal-to-Noise Ratio (PSNR), Structural Similari…”
MethodInfluentialColor Learning for Image Compression
Prativadibhayankaram, S., Richter, T., Sparenberg, H., et al. (2023) · Proceedings - International Conference on Image Processing · doi:10.1109/icip49359.2023.10222731 · Cited by 7
“Some of the most commonly used metrics for distortion are mean squared error (MSE), peak signal to noise ratio (PSNR), and multi scale structural similarity index (MS-SSIM) [4].”
MethodInfluentialWang, Y., Huang, Q., Liu, J., Jiang, C., & Shang, M. (2023) · Multimedia Tools and Applications · doi:10.1007/s11042-023-16607-z · Cited by 9
TCM-Net: Mixed Global–Local Learning for Salient Object Detection in Optical Remote Sensing Images
He, J., Zhao, L., Hu, W., et al. (2023) · Remote Sensing · doi:10.3390/rs15204977 · Cited by 7
That's What I Said: Fully-Controllable Talking Face Generation
Jang, Y., Rho, K., Woo, J., Lee, H., Park, J., Lim, Y., Kim, B.-Y., & Chung, J. S. (2023) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3581783.3612587 · Cited by 13
“For MEAD dataset, our framework shows the best performance on MS-SSIM, PSNR, and LMD and comparable results to the Wav2Lip model in terms of SSIM.”
MethodLoss functions in machine learning for seismic random noise attenuation
Jun, H., & Kim, H. (2023) · Geophysical Prospecting · doi:10.1111/1365-2478.13449 · Cited by 7
“Equation (7) shows the MS-SSIM in the case of 𝛽 𝑗 = 𝛾 𝑗 (Wang et al., 2003).”
MethodAdaptive Single Low-Light Image Enhancement by Fractional Stretching in Logarithmic Domain
Trongtirakul, T., Agaian, S. С., & Wu, S. (2023) · IEEE Access · doi:10.1109/access.2023.3344534 · Cited by 7
“In this paper, several FR-IQA metrics, namely the multiscale structural similarity (MS-SSIM) [39], the lightness order error (LOE) [30], and the peak single-to-noise ratio (PSNR) [38], MOS [40], were employed to evaluate the performance of various low-light image enhancement methods.”
MethodResidual Attention Augmented U-Shaped Network for One-Bit SAR Image Restoration
Guo, L. B., Dong, Y., & Dong, C. (2024) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2024.3357812 · Cited by 8
“Compared with the single-scale SSIM index, the MSSSIM incorporates evaluation results of image details at different resolutions based on richer structural information [70], which is defined as where exponents of α G , β g , and γ g are also used to adjust the relative importance of three components.”
BackgroundMethodZhang, X., Bernd Hamann, A., Wang, D., et al. (2024) · IEEE Transactions on Consumer Electronics · doi:10.1109/tce.2024.3386773 · Cited by 7
“…loss functions include L 1 reconstruction loss for hole regions L hole and valid regions L valid , Multi-scale Structural SIMilarity (MS-SSIM) loss [31] L ms and Total Variation (TV) loss L TV , i.e., where M , M gt , M com and M m denote the image generated by FMGDN, the ground truth, the…”
BackgroundMethodSon, Y., Jeong, S., Hong, Y., Lee, J., Jeon, B., Choi, H., Kim, J.-H., & Shim, H. (2024) · Bioengineering · doi:10.3390/bioengineering11090944 · Cited by 8
“In the field of image enhancement or denoising, GAN-based solutions with adversarial training are better at constructing manifolds closer to the target data than MSE-based solutions and focus on perceptual quality rather than optimizing only for quantitative metrics [11].”
Background2D NO-REFERENCE VIDEO QUALITY MODEL DEVELOPMENT AND 3D VIDEO TRANSMISSION QUALITY
Brunnstrom, K. E., Sedano, I., Wang, K., Barkowsky, M., Kihl, M., Andrén, B., Callet, P. L., Aurelius, A., Ab, A., & France, I. N. (2012) · Cited by 7
Dumić, E., & Grgić, S. (2013) · International Symposium ELMAR (ELMAR) · Cited by 7
Design of Objective Video Quality Metrics Using Spatial and Temporal Informations
Regis, C. D. M., Oliveira, Í. D. P., Cardoso, J. V. D. M., & de Alencar, M. S. (2015) · IEEE Latin America Transactions · doi:10.1109/tla.2015.7069106 · Cited by 7
Reversible De-Identification for lossless image compression using Reversible Watermarking
Farrugia, R. A. (2014) · International Convention on Information and Communication Technology, Electronics and Microelectronics · doi:10.1109/mipro.2014.6859760 · Cited by 7
“The proposed Threshold Selection method is based on the observation that different sub-bands provide dissimilar levels of distortions [11].”
MethodPerformance optimizations for PatchMatch-based pixel-level multiview inpainting
Lu, S.-P., Ceulemans, B., Munteanu, A., et al. (2013) · International Conference on D Imaging · doi:10.1109/ic3d.2013.6732089 · Cited by 7
“Here we employ Peak Signal-to-noise Ratio (PSNR), Structural Similarity (SSIM) [14] and Multi-scale Structural Similarity (MSSSIM) [15] to evaluate the inpainting result.”
MethodSuper-Resolving IC Images With an Edge-Preserving Bayesian Framework
Wang, Z., Yang, H. S., Li, W., & Yin, Z. (2014) · IEEE Transactions on Semiconductor Manufacturing · doi:10.1109/tsm.2013.2293581 · Cited by 7
“Referring to the investigation to the methods of image quality assessment in [37], the structural similarity (SSIM) [38], the multi-scale structural similarity (MS-SSIM) [39], the naturalness Image quality evaluator (NIQE) [40] and the blind or no-reference image spatial quality evaluator (BRISQUE) [37] are very effec…”
MethodSelf-citationFull-reference video quality assessment on high-definition video content
Wulf, S., & Zölzer, U. (2012) · International Conference on Signal Processing and Communication Systems · doi:10.1109/icspcs.2012.6507948 · Cited by 7
A structural difference based image clutter metric with brain cognitive model constraints
Xu, D., Shi, Z., & Luo, H. (2012) · Infrared Physics & Technology · doi:10.1016/j.infrared.2012.11.005 · Cited by 8
ViMSSIM: from image to video quality assessment
Vũ, C. T., & Deshpande, S. (2012) · MoVid '12 · doi:10.1145/2151677.2151679 · Cited by 9
A non-reference measure for objective edge map evaluation
Nercessian, S. C., Panetta, K. A., & Agaian, S. С. (2009) · Conference proceedings/Conference proceedings - IEEE International Conference on Systems, Man, and Cybernetics · doi:10.1109/icsmc.2009.5346779 · Cited by 7
“In such methods, high order PDEs are designed to restore smooth regions as well as thin structures [12] [13].”
MethodQuaternionic wavelet coefficients modeling for a Reduced-Reference metric
Traoré, A., Carré, P., & Olivier, C. (2015) · Signal Processing Image Communication · doi:10.1016/j.image.2015.06.003 · Cited by 8
Blind Quality Metric for Multidistortion Images Based on Cartoon and Texture Decomposition
Zhang, F., & Roysam, B. (2016) · IEEE Signal Processing Letters · doi:10.1109/lsp.2016.2594166 · Cited by 7
A reduced reference video quality assessment of H.264 and mpeg2 codec based on sharpness metric
Amor, M. B., Larabi, M., Kammoun, F., & Masmoudi, N. (2016) · Journal of the Society for Information Display · doi:10.1002/jsid.517 · Cited by 7
Fast algorithm for image denoising with different boundary conditions
Zhang, X., Shi, Y., Pang, Z., et al. (2017) · Journal of the Franklin Institute · doi:10.1016/j.jfranklin.2017.04.011 · Cited by 7
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.