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 1,501–1,600 of 6,047
Task-Driven Video Compression for Humans and Machines: Framework Design and Optimization
Yi, X., Wang, H., Kwong, S., & Kuo, C. ‐. J. (2022) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2022.3233245 · Cited by 34
“Peak signal-to-noise ratio (PSNR) and multi-scale structural similarity (MS-SSIM) [45] are used to evaluate the distortion between the reconstructed frames and the raw frames.”
MethodHigh Dynamic Range Image Quality Assessment Based on Frequency Disparity
Liu, Y., Ni, Z., Wang, S., Wang, H., & Kwong, S. (2023) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2023.3237702 · Cited by 28
“To illustrate the superiority of the proposed LGFM, several classic and state-of-the-art IQA metrics are adopted for comparison, including HDR-VDP-3 [9], HDR-VQM [11], PSNR, SSIM [13], FSIM [16], VIF [15], GMSD [14], ESIM [20], GFM [21], and MSSIM [31], where the first two algorithms are specifically designed for HDR…”
MethodLearning to Super-Resolve Blurry Images With Events
Yu, L., Wang, B., Zhang, X., Zhang, H., Yang, W. T., Liu, J., & Xia, G.-S. (2023) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2023.3240397 · Cited by 30
“2), the DSL scheme brings remarkable improvements in terms of PSNR and SSIM on both GoPro and HQF datasets, i.e. , 0 .”
MethodInfluentialAutomatic Quantitative Coronary Analysis Based on Deep Learning
Liu, X., Wang, X., Chen, D., et al. (2023) · Applied Sciences · doi:10.3390/app13052975 · Cited by 22
The Bjøntegaard Bible Why Your Way of Comparing Video Codecs May Be Wrong
Herglotz, C., Och, H., Meyer, A. P., Ramasubbu, G., Eichermüller, L., Kränzler, M., Brand, F., Fischer, K., Nguyen, D. T., Regensky, A., et al. (2024) · IEEE Transactions on Image Processing · doi:10.1109/tip.2023.3346695 · Cited by 34
Multi-view aggregation transformer for no-reference point cloud quality assessment
Mu, B., Shao, F., Chai, X., Liu, Q., Chen, H., & Jiang, Q. (2023) · Displays · doi:10.1016/j.displa.2023.102450 · Cited by 23
An Efficient and Robust Underwater Image Compression Scheme Based on Autoencoder
Liu, J., Yuan, F., Xue, C., et al. (2023) · IEEE Journal of Oceanic Engineering · doi:10.1109/joe.2023.3249243 · Cited by 24
“For better quantitative comparison and analysis, we use multiscale structural similarity index measure (MS-SSIM) [29] to select the optimal one from several reconstructed images.”
MethodElastic Seismic Imaging Enhancement of Sparse 4C Ocean-Bottom Node Data Using Deep Learning
Cheng, S., Shi, X., Mao, W., et al. (2023) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2023.3275614 · Cited by 22
“Finally, MS-SSIM evaluation comes from combining measurements at different scales as follows: MS-SSIM(L , O) = [lM(L , O)]αM · M∏ j=1 [cM(L , O)]β j [sM(L , O)]γ j (4) with l(L , O) = 2µLµO + C1 µ2L + µ 2 O + C1 c(L , O) = 2σLσO + C2 σ 2L + σ 2 O + C2 s(L , O) = σL O + C3 σLσO + C3 (5) where µL and µO denote the mean…”
MethodInfluentialLessons from a Space Lab: An Image Acquisition Perspective
Pauly, L., Jamrozik, M. L., del Castillo, M. O., Borgue, O., Singh, I. P., Makhdoomi, M. R., Christidi-Loumpasefski, O.-O., Gaudillière, V., Martínez, C., Rathinam, A., et al. (2023) · International Journal of Aerospace Engineering · doi:10.1155/2023/9944614 · Cited by 26
Wu, G., Fu, H., Liu, J., Ma, L., Fan, X., & Liu, R. (2024) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v38i6.28413 · Cited by 35
Perceptual experience of time-varying video quality
Rehman, A., & Wang, Z. (2013) · International Conference on Quality of Multimedia Experience (QoMEX) · doi:10.1109/qomex.2013.6603240 · Cited by 22
Self-citationBackwards Compatible Coding of High Dynamic Range Images with JPEG
Richter, T. (2013) · DCC · doi:10.1109/dcc.2013.24 · Cited by 22
Kim, K. J., Kim, B., Mantiuk, R., et al. (2010) · IEEE Transactions on Medical Imaging · doi:10.1109/tmi.2010.2049655 · Cited by 22
Perceptual screen content image quality assessment and compression
Wang, S., Gu, K., Zeng, K., Wang, Z., & Lin, W. (2015) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2015.7351037 · Cited by 21
“We verify the quality assessment method by comparing its performance with PSNR, SSIM [2], MS-SSIM [14], IW-SSIM[4], GSIM [9], FSIM [5], GMSD [15], VSI [16] and VIF [3] using the newly proposed SIQAD database [7].”
MethodSelf-citationObjective quality metric for 3D virtual views
Farid, M. S., Lucenteforte, M., & Grangetto, M. (2015) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2015.7351499 · Cited by 21
“Moreover, we compare SIQE with 6 widely used full reference 2D image quality metrics - SSIM [33], FSIM [34], MSSIM [35], iwSSIM [36], Peak Signal to Noise Ratio (PSNR), and UQI [37].”
MethodInfluentialA Detail-Based Method for Linear Full Reference Image Quality Prediction
Claudio, E. D. D., & Jacovitti, G. (2017) · IEEE Transactions on Image Processing · doi:10.1109/tip.2017.2757139 · Cited by 26
“socalled structural information, defined through the local normalized Mean Square Error (MSE) between the test and the reference images [12]. Variants of the SSIM include MultiScale analysis (MS-SSIM) [13] or image gradient analysis, adopted in the Feature Similarity (FSIM) [14] and in the Gradient Magnitude Similarity…”
MethodBosse, S., Becker, S., Müller, K., Samek, W., & Wiegand, T. (2018) · Digital Signal Processing · doi:10.1016/j.dsp.2018.12.005 · Cited by 23
“Two main strategies are classically distinguished for FR IQMs [6]: whereas bottom-up approaches explicitly emulate the human visual system (HVS) [5, 7, 8], top-down approaches model hypothesized abstract processing properties of the HVS from a signal processing perspective [9, 10, 11, 12].”
BackgroundMethodScreen Content Image Quality Assessment With Edge Features in Gradient Domain
Wang, R., Yang, H., Pan, Z., Huang, B.-X., & Hou, G.-J. (2019) · IEEE Access · doi:10.1109/access.2018.2889992 · Cited by 19
iTM-Net: Deep Inverse Tone Mapping Using Novel Loss Function Considering Tone Mapping Operator
Kinoshita, Y., & Kiya, H. (2019) · IEEE Access · doi:10.1109/access.2019.2919296 · Cited by 22
“73560 VOLUME 7, 2019 VOLUME 7, 2019 73561 B. RESULTS Figures 3 and 4 summarize quantitative evaluation results as box plots for the 44 images in terms of HDR-VDP and PU encoding+MS-SSIM, respectively.”
MethodInfluentialVoronoi-based Objective Quality Metrics for Omnidirectional Video
Croci, S., Ozcinar, C., Zerman, E., Cabrera, J., & Smolić, A. (2019) · International Conference on Quality of Multimedia Experience (QoMEX) · doi:10.1109/qomex.2019.8743345 · Cited by 24
“Between these two metrics, VMAF is slightly better than MS-SSIM for both representation formats.”
BackgroundMethodInfluentialHigh-Fidelity Face Sketch-To-Photo Synthesis Using Generative Adversarial Network
Chao, W., Chang, L., Wang, X., et al. (2019) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2019.8803549 · Cited by 22
“We adopt standard evaluation metrics such as Inception Score [14], FID (Fréchet Inception Distance) [15], PSNR, SSIM [16] and MS-SSIM [17].”
MethodInfluentialDynamic Range Expansion Using Cumulative Histogram Learning for High Dynamic Range Image Generation
Jang, H., Bang, K., Jang, J., & Hwang, D. (2020) · IEEE Access · doi:10.1109/access.2020.2975857 · Cited by 22
“2 [44], [45], PU-SSIM(Y) [46], [47], PU-MS-SSIM [48], PU-PSNR, PU-RMSE, and deltaE1976 [37].”
Motion-supervised Co-Part Segmentation
Siarohin, A., Roy, S., Lathuilière, S., Tulyakov, S., Ricci, E., & Sebe, N. (2021) · Proceedings - International Conference on Pattern Recognition/Proceedings/International Conference on Pattern Recognition · doi:10.1109/icpr48806.2021.9412520 · Cited by 40
“In addition, this loss is computed at several resolutions, similarly to MS-SSIM [39], [40].”
MethodA Local Flatness Based Variational Approach to Retinex
Tang, M., Xie, F., Zhang, R., Jiang, Z., & Bovik, A. C. (2020) · IEEE Transactions on Image Processing · doi:10.1109/tip.2020.2999858 · Cited by 24
“We also calculated the RMSE (root mean square error), MAE (mean absolute error), SSIM (structural similarity) [46], MSSSIM (multi-scale structural similarity) [47] and FSIM (feature similarity) [48] of the six methods on the results computed on the 10 synthetic images.”
MethodSelf-citationPerceptual similarity between color images using fuzzy metrics
Grečova, S., & Morillas, S. (2015) · Journal of Visual Communication and Image Representation · doi:10.1016/j.jvcir.2015.04.003 · Cited by 23
“This similarity measure is extended to the Multiscale Structural Similarity Index (MSSIM) in [15].”
MethodLiu, S., Lan, X., Chen, W., Zhang, Z., & Qiu, C. (2024) · IEEE Sensors Journal · doi:10.1109/jsen.2024.3426274 · Cited by 25
“Furthermore, we devise a content loss function consisting of multiscale structural similarity (MS-SSIM) [23] loss and spatial frequency (SF) [24] loss, effectively maintaining detailed information and salient features of source images.”
BackgroundMethodsDMCM—A Semantic Digital Modulation Constellation Mapping Scheme for Semantic Communication
Teng, L., An, W., Dong, C., & Xu, X. (2025) · IEEE Internet of Things Journal · doi:10.1109/jiot.2025.3545667 · Cited by 28
Multimodal Medical Image Fusion Using a Progressive Parallel Strategy Based on Deep Learning
Peng, P., & Luo, Y. (2025) · Electronics · doi:10.3390/electronics14112266 · Cited by 24
Li, M., Zhang, Y., Qu, W., Zhu, W., & Ke, Y. (2025) · Composites Science and Technology · doi:10.1016/j.compscitech.2025.111361 · Cited by 25
Visual attention based image quality assessment
Guo, A., Zhao, D., Liu, S., Fan, X., & Gao, W. (2011) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2011.6116375 · Cited by 22
Ground truth study on fractal dimension of color images of similar texture
Nayak, S. R., Khandual, A., & Mishra, J. (2017) · Journal of the Textile Institute · doi:10.1080/00405000.2017.1418710 · Cited by 22
Learned Resolution Scaling Powered Gaming-as-a-Service at Scale
Chen, H., Lu, M., Ma, Z., Zhang, X., Xu, Y., Shen, Q., & Zhang, W. (2020) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2020.2985538 · Cited by 24
“Note that distortion evaluation in traditional BD-Rate was replaced by the Multi-Scale Structural Similarity (MS-SSIM) [18] for its close correlation with perceptual quality as the QoE index.”
MethodInfluentialSynthesis and Edition of Ultrasound Images via Sketch Guided Progressive Growing GANS
Liang, J., Yang, X., Li, H., et al. (2020) · IEEE International Symposium on Biomedical Imaging (ISBI) · doi:10.1109/isbi45749.2020.9098384 · Cited by 21
“Table 1 further lists the numerical evaluation metrics, including Freshet Inception Distance (FID) [14], Kernel Inception Distance (KID) [15] and multi-scale structural similarity (MS-SSIM) [16].”
BackgroundFall detection based on fused saliency maps
Li, H., Li, C., & Ding, Y. (2020) · Multimedia Tools and Applications · doi:10.1007/s11042-020-09708-6 · Cited by 22
“Specially, Mean Structural SIMilarity (MSSIM) [39] is introduced to learn the weights.”
BackgroundDeep learning for simultaneous seismic image super-resolution and denoising
Li, J., Wu, X., & Hu, Z. (2020) · Second International Meeting for Applied Geoscience & Energy · doi:10.1190/segam2020-3426412.1 · Cited by 21
Lossy Medical Image Compression using Residual Learning-based Dual Autoencoder Model
Mishra, D., Singh, S. K., & Singh, R. K. (2020) · IEEE Uttar Pradesh Section International Conference on Electrical, Electronics and Computer Engineering (UPCON) · doi:10.1109/upcon50219.2020.9376417 · Cited by 21
“The proposed algorithm has been evaluated for the quality of reconstructed images with the help of peak signal-to-noise ratio (PSNR) [22] and multi-structural similarity (MS-SSIM) [23].”
MethodPhotogrammetric Digital Surface Model Reconstruction in Extreme Low-Light Environments
Roncella, R., Bruno, N., Diotri, F., Thoeni, K., & Giacomini, A. (2021) · Remote Sensing · doi:10.3390/rs13071261 · Cited by 23
“Several modifications and integrations of the original SSIM index have been proposed in the last decade (e.g., complex wavelet SSIM [43], MS-SSIM [44]) to increase its performances or deal with specific issues, such as image scale or rotation.”
BackgroundSymmetrical lattice generative adversarial network for remote sensing images compression
Zhao, S., Yang, S., Gu, J., Liu, Z., & Feng, Z. (2021) · ISPRS Journal of Photogrammetry and Remote Sensing · doi:10.1016/j.isprsjprs.2021.03.009 · Cited by 25
Discriminator-free Generative Adversarial Attack
Lu, S., Xian, Y., Yan, K., Hu, Y., Sun, X., Guo, X., Huang, F., & Zheng, W.-S. (2021) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3474085.3475290 · Cited by 24
Reference-free learning-based similarity metric for motion compensation in cone-beam CT
Huang, H., Siewerdsen, J. H., Zbijewski, W. B., Weiss, C. R., Unberath, M., Ehtiati, T., & Sisniega, A. (2022) · Physics in Medicine and Biology · doi:10.1088/1361-6560/ac749a · Cited by 27
Learning to See Through With Events
Yu, L., Zhang, X., Liao, W., Yang, W. T., & Xia, G.-S. (2022) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2022.3227448 · Cited by 26
“structural similarity (SSIM) [52] as the ground truth.”
BackgroundInfluentialWang, J., Xu, G., Li, C., Gao, G., & Wu, Q. W. (2022) · IEEE Sensors Journal · doi:10.1109/jsen.2022.3229031 · Cited by 22
“Multiscale SSIM [47] is originally used to measure SSIM between the prediction result and ground truth at different resolutions.”
MethodAfnan, A., Ullah, F., Yaseen, Y., Lee, J., Jamil, S., & Kwon, O.-J. (2023) · Italian National Conference on Sensors · doi:10.3390/s23031297 · Cited by 19
“The objective IQMs, named FSIM, MS-SSIM, and IW-SSIM, show the least standard deviations with a close range of values that best guarantee the visually or near-visually lossless compression of the images.”
MethodInfluentialImage Completion with Heterogeneously Filtered Spectral Hints
Xu, X., Navasardyan, S., Tadevosyan, V., Sargsyan, A., Mu, Y., & Shi, H. (2023) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv56688.2023.00457 · Cited by 35
“The LPIPS, PSNR and SSIM scores agree with our FID score.”
MethodZheng, Q., Zheng, L., Bai, Y., Liu, H., Deng, J., & Li, Y. (2023) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2023.3260825 · Cited by 30
“Here, we use the SSIM [30] loss originally used for image quality assessment to capture the structural topology of salient objects in optical RSIs.”
MethodA curvature-driven cloud removal method for remote sensing images
Yu, X., Pan, J. H., Wang, M., et al. (2023) · Geo-spatial Information Science · doi:10.1080/10095020.2023.2189462 · Cited by 21
“In the quantitative analysis, MSSIM, MS-SSIM, PSNR, CC and CE values are calculated to evaluate the cloud removal results of different buffer areas.”
InfluentialEnd-To-End Compression for Surveillance Video With Unsupervised Foreground-Background Separation
Zhao, Y. Q., Luo, D., Wang, F., Gao, H., Ye, M., & Zhu, C. (2023) · IEEE Transactions on Broadcasting · doi:10.1109/tbc.2023.3280039 · Cited by 21
CA-GAN: the synthesis of Chinese art paintings using generative adversarial networks
Chen, Z., & Zhang, Y. (2023) · The Visual Computer · doi:10.1007/s00371-023-03115-2 · Cited by 25
“Under such circumstances, we apply MS-SSIM [31] to remedy the above deficiency, by paying more attention to the local features of an image:”
Classifier-guided multi-style tile image generation method
Lu, J., Shi, M., Song, C., Zhao, W., Xi, L., & Emam, M. (2023) · Journal of King Saud University - Computer and Information Sciences · doi:10.1016/j.jksuci.2023.101899 · Cited by 21
“The larger MS-SSIM value indicates a more significant similarity texture structure of the image.”
InfluentialA single-image GAN model using self-attention mechanism and DenseNets
Yıldız, E., Yüksel, M. E., & Sevgen, S. (2024) · Neurocomputing · doi:10.1016/j.neucom.2024.127873 · Cited by 25
Nonparametric Multiscale Blind Estimation of Intensity-Frequency-Dependent Noise
Colom, M., Lebrun, M., Buades, A., & Morel, J. (2015) · IEEE Transactions on Image Processing · doi:10.1109/tip.2015.2438537 · Cited by 22
Image haze removal: Status, challenges and prospects
Wu, D., Zhu, Q., Wang, J., et al. (2014) · IEEE International Conference on Information Science and Technology (ICIST) · doi:10.1109/icist.2014.6920524 · Cited by 22
Details preservation inspired blind quality metric of tone mapping methods
Gu, K., Zhai, G., Liu, M., Yang, X., & Zhang, W. (2014) · IEEE International Symposium on Circuits and Systems proceedings · doi:10.1109/iscas.2014.6865186 · Cited by 22
Towards an Evaluation of Denoising Algorithms with Respect to Realistic Camera Noise
Seybold, T., Keimel, C., Knopp, M., & Stechele, W. (2013) · IEEE International Symposium on Multimedia · doi:10.1109/ism.2013.39 · Cited by 23
“However, VIF, IFC, PSNR-HVS, VSNR and MSSIM showed lower results for the noisy images when demosaicking is included compared to the uncorrelated noise, and thus matched this important tendency shown by the MOS.”
BackgroundInfluentialStructural Similarity Measure to Assess Improvement by Noise in Nonlinear Image Transmission
Rousseau, D., Delahaies, A., & Chapeau‐Blondeau, F. (2009) · IEEE Signal Processing Letters · doi:10.1109/lsp.2009.2031734 · Cited by 22
“Although still in an early stage, the SSIM index, from the recent results [1], [20], [2], [21], [22], [18], [23], [24], [19], is gradually emerging as a specially useful metric for image quality, providing a specific intermediate approach exhibiting”
BackgroundMethodFast inverse tone mapping with Reinhard's global operator
Kinoshita, Y., Shiota, S., & Kiya, H. (2017) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp.2017.7952501 · Cited by 21
Using multiple spatio-temporal features to estimate video quality
Freitas, P. G., Akamine, W. Y. L., & Farias, M. C. Q. (2018) · Signal Processing Image Communication · doi:10.1016/j.image.2018.02.010 · Cited by 24
“In this work, we use the mean of the SSIM index map to quantify the quality difference between of u and v, at different scales [35].”
MethodYang, J., Liu, J., Jiang, B., & Lu, W. B. (2018) · Signal Processing · doi:10.1016/j.sigpro.2018.07.006 · Cited by 21
Deep One-Class Classification Using Intra-Class Splitting
Schlachter, P., Liao, Y., & Yang, B. (2019) · IEEE Data Science Workshop (DSW) · doi:10.1109/dsw.2019.8755576 · Cited by 35
Artificial neural network for Slice Encoding for Metal Artifact Correction (SEMAC) MRI
Seo, S., Do, W., Luu, H. M., Kim, K. H., Choi, S. H., & Park, S. (2019) · Magnetic Resonance in Medicine · doi:10.1002/mrm.28126 · Cited by 23
Ledesma‐Gil, G., Fernández-Avellaneda, P., & Spaide, R. F. (2020) · Retina · doi:10.1097/iae.0000000000002866 · Cited by 21
Efficient structurally-strengthened generative adversarial network for MRI reconstruction
Zhou, W., Du, H., Mei, W., & Fang, L. (2020) · Neurocomputing · doi:10.1016/j.neucom.2020.09.008 · Cited by 22
“According to [25], MS-SSIM outperforms SSIM under an appropriate parameter settings.”
MethodAn Infrared and Visible Image Fusion Algorithm Based on LSWT-NSST
Li, J., Li, B., & Jiang, Y. (2020) · IEEE Access · doi:10.1109/access.2020.3028088 · Cited by 21
“MS-SSIM [46] is an indicator based on the human visual sys-tem.”
BackgroundMethodData-Driven Transform-Based Compressed Image Quality Assessment
Zhang, X., Kwong, S., & Kuo, C. ‐. J. (2020) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2020.3041639 · Cited by 24
CloTH-VTON+: Clothing Three-Dimensional Reconstruction for Hybrid Image-Based Virtual Try-ON
Minar, M. R., Tuan, T. T., & Ahn, H. (2021) · IEEE Access · doi:10.1109/access.2021.3059701 · Cited by 28
Deep Tone-Mapping Operator Using Image Quality Assessment Inspired Semi-Supervised Learning
Guo, C., & Jiang, X. (2021) · IEEE Access · doi:10.1109/access.2021.3080331 · Cited by 21
“Finally, SF ( X , Y ) of all different scales are calculated into structure fidelity part of TMQI using same coefficients as MS-SSIM [52].”
MethodImperceptible reversible watermarking of radiographic images based on quantum noise masking
Wei, P., Bouslimi, D., Karasad, M., et al. (2018) · Computer Methods and Programs in Biomedicine · doi:10.1016/j.cmpb.2018.03.011 · Cited by 21
Image quality assessment using contourlet transform
Liu, M., & Yang, X. (2009) · Optical Engineering · doi:10.1117/1.3241996 · Cited by 21
Global-Local Face Upsampling NetworkPreprint
Tuzel, O., Taguchi, Y., & Hershey, J. R. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1603.07235 · Cited by 46
Invisible Steganography via Generative Adversarial NetworksPreprint
Zhang, R., Dong, S., & Liu, J. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1807.08571 · Cited by 21
“Inspired by their work, we introduce the structure similarity index (SSIM) [17] and its variant, the multi-scale structure similarity index (MS-SSIM) [18] into our metric.”
MethodComputationally Efficient Neural Image CompressionPreprint
Johnston, N., Eban, E., Gordon, A., & Ballé, J. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1912.08771 · Cited by 75
“te-distortion performance with respect to classic image compression codecs [24,12,10,8] until they were first match or exceeded in Mean-Squared Error (MSE) or Multiscale Structural Similarity(MS-SSIM) [25] in [7]. Despite this success in the field of neural image compression, neural methods have not be widely deployed.…”
BackgroundMethodInfluentialObjective quality assessment of medical images and videos: review and challenges
Rodrigues, R., Lévêque, L., Gutiérrez, J., Jebbari, H., Outtas, M., Zhang, L., Chetouani, A., Al-Juboori, S., Martini, M. G., & Pinheiro, A. (2024) · Multimedia Tools and Applications · doi:10.1007/s11042-024-20292-x · Cited by 40
“…used in both studies, but a larger set of metrics was used in [5], including, for example, PSNR-HVS and PSNR-HVS-M [44], MSE, Multi-scale SSIM (MS-SSIM) [45], and two NR metrics - Natural Image Quality Evaluator (NIQE) [46] and Blind/Referenceless Image Spatial Quality Ele-vator (BRISQUE) [47].”
Schaudt, D., Späte, C., von Schwerin, R., Reichert, M., von Schwerin, M., Beer, M., & Kloth, C. (2023) · Bioengineering · doi:10.3390/bioengineering10121421 · Cited by 25
“Fidelity and diversity can be measured by the Fréchet Inception Distance (FID) [58] and the Multi-Scale Structural Similarity Index (MS-SSIM) [59], respectively.”
MethodYang, Z., Shen, D. G., Chan, K. W. Y., & Huang, J. (2024) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2024.3404225 · Cited by 23
Liao, X., Wei, X., Zhou, M., Wong, H., & Kwong, S. (2025) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2025.3527004 · Cited by 29
Hanhart, P., & Ebrahimi, T. (2012) · 3DTV-Conference: The True Vision - Capture, Transmission and Display of 3D Video · doi:10.1109/3dtv.2012.6365478 · Cited by 20
“MS-SSIM: Multi-Scale Structural Similarity Index [6], 4.”
Increasing User Perceived Quality by Selective Load Balancing of Video Traffic in Wireless Networks
Hava, A., Ghamri-Doudane, Y. M., Muntean, G., et al. (2015) · IEEE Transactions on Broadcasting · doi:10.1109/tbc.2015.2400811 · Cited by 20
Visual structural degradation based reduced-reference image quality assessment
Wu, J., Lin, W., Fang, Y., Li, L., Shi, G., & Issac Niwas, S. (2016) · Signal Processing Image Communication · doi:10.1016/j.image.2016.05.008 · Cited by 21
Blind Stereo Quality Assessment Based on Learned Features From Binocular Combined Images
Karimi, M., Nejati, M., Soroushmehr, S. M. R., Samavi, S., Karimi, N., & Najarian, K. (2017) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2017.2699082 · Cited by 20
“792 MS-SSIM [2] 0.”
Yılmaz, Ç. Ş., Yılmaz, V., & Güngör, O. (2020) · International Journal of Remote Sensing · doi:10.1080/01431161.2019.1711244 · Cited by 20
“The PCA result, in the site 1, got the RMSE, RASE, ERGAS, IW-SSIM, UIQI, MS-SSIM, SSIM, SID, SAM and r values of 75.296, 4.129, 9.260, 0.788, 0.384, 0.878, 0.841, 0.215, 5.223 and 0.926, respectively.”
InfluentialNSCT-Based Robust and Perceptual Watermarking for DIBR 3D Images
Nam, S.-H., Mun, S.-M., Ahn, W., Kim, D., Yu, I.-J., Kim, W., & Lee, H.-K. (2020) · IEEE Access · doi:10.1109/access.2020.2994966 · Cited by 21
“…the peak-signal-to-noise ratio (PSNR) and structural similarity (SSIM) [50], which are most actively used for the evaluation criteria of image quality, we analyzed the quality of watermarked images using the following IQA metrics: feature similarity (FSIM) [51] and multi-scale SSIM (MS-SSIM) [52].”
MethodReference Guided Face Component Editing
Deng, Q., Cao, J., Liu, Y., Chai, Z., Li, Q., & Sun, Z. (2020) · International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2020/70 · Cited by 24
“…methods [He et al., 2019; Wu et al., 2019], we leverage Fréchet Inception Distance (FID, lower value indicates better quality) [Heusel et al., 2017] and Multi-Scale Structural SIMilarity (MS-SSIM, higher value indicates better quality) [Wang et al., 2003] to evaluate the performance of our model.”
MethodSandic-Stankovic, D. D., Kukolj, D., & Callet, P. L. (2022) · IEEE Transactions on Image Processing · doi:10.1109/tip.2021.3139238 · Cited by 21
A Robust Model for MODIS and Landsat Image Fusion Considering Input Noise
Tan, Z., Gao, M., Yuan, J. J., et al. (2022) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2022.3145086 · Cited by 20
“The vision loss can be calculated with the multiscale structural similarity (MS-SSIM) [47], as shown in the following equation : where I denotes the multidimensional tensor of ones. l i ( x , y ) , c i ( x , y ) , and s i ( x , y ) are the quantified luminance, contrast, and structure of predicted…”
Saincher, R. R., Kumar, S., Gopalkrishna, P., Maithri, M., & Sherigar, P. (2022) · Heliyon · doi:10.1016/j.heliyon.2022.e09262 · Cited by 35
“IQA offers a simplistic approach, close to human perception, for evaluating any image [29].”
BackgroundMorphology-Preserving Autoregressive 3D Generative Modelling of the Brain
Tudosiu, P.-D., Pinaya, W. H. L., Graham, M. S., Borges, P., Fernández, V., Yang, D., Appleyard, J., Novati, G., Mehra, D., Vella, M., et al. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-16980-9_7 · Cited by 35
“Lastly, to estimate the diversity of the generated images, MS-SSIM is being used in a pair-wise fashion between the generated synthetic samples as in [8, 9].”
InfluentialViewport-Sphere-Branch Network for Blind Quality Assessment of Stitched 360° Omnidirectional Images
Tian, C., Shao, F., Chai, X., Jiang, Q., Xu, L., & Ho, Y. (2022) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2022.3225172 · Cited by 25
“One obvious shortcoming for these methods is that there is some overlapping among constituent images, thus, the pixel-wise FR methods [10], [11] cannot be directly used for measuring stitching distortion.”
MethodBlind Image Despeckling Using a Multiscale Attention-Guided Neural Network
Guo, Y., Lu, Y., Liu, R. W., & Zhu, F. (2023) · IEEE Transactions on Artificial Intelligence · doi:10.1109/tai.2023.3235342 · Cited by 22
“More details on the definitions of L MS-SSIM and L 1 can be found in [43] and [44].”
BackgroundRadhakrishna, C. G., & Ciuciu, P. (2023) · Bioengineering · doi:10.3390/bioengineering10020158 · Cited by 20
“Loss, Gradients and Optimizer The reconstruction error used as loss function Lr (between the reference MR image x and its reconstruction x̂) in this study was inspired by [30] and is defined as a weighted sum of `1, `2 and the multi-scale structural similarity index (S) [31]:”
MethodSpectral Clustering Super-Resolution Imaging Based on Multispectral Camera Array
Huang, F., Chen, Y., Wang, X., Wang, S., & Wu, X. (2023) · IEEE Transactions on Image Processing · doi:10.1109/tip.2023.3242589 · Cited by 21
“The quantitative methods include PSNR, MultiSSIM [71], and BRISQUE [72].”
Full-Reference Image Quality Expression via Genetic Programming
Bakurov, I., Buzzelli, M., Schettini, R., Castelli, M., & Vanneschi, L. (2023) · IEEE Transactions on Image Processing · doi:10.1109/tip.2023.3244662 · Cited by 13
El‐Fattah, I. A., Ali, A. M., El‐Shafai, W., Taha, T. E., & El‐Samie, F. E. A. (2023) · Optical and Quantum Electronics · doi:10.1007/s11082-022-04432-x · Cited by 21
“Also, different methods have been applied to measure the performance of the SR models, including PSNR, SSIM, and MS-SSIM.”
MethodInfluentialBourbia, S., Karine, A., Chetouani, A., Hassouni, M. E., & Jridi, M. (2023) · IEEE Access · doi:10.1109/access.2023.3247191 · Cited by 21
“Torlig et al. [35] projected the voxelized PC into six ortho-graphic viewpoints and employed 2D objective quality meth-ods, including Peak Signal-to-Noise Ratio (PSNR) [36], Information Fidelity in Pixel domain (VIFP) [37], Structural Similarity Index (SSIM) [27] and Multi Scale Structural Similarity Index (MS-SSIM) […”
BackgroundXu, Y. K. T., Graves, A. R., Coste, G. I., Huganir, R. L., Bergles, D. E., Charles, A. S., & Sulam, J. (2023) · Nature Methods · doi:10.1038/s41592-023-01871-6 · Cited by 22
“Combining this with other losses, such as the multi-scale structural similarity index (MS-SSIM) (Wang et al., 2003), is certainly possible and might provide further improvements.”
BackgroundChen, J., Huang, C.-K., Delgado, J. F., & Misra, S. (2023) · Computational Geosciences · doi:10.1007/s10596-023-10243-0 · Cited by 23
“MS-SSIM is conducted at different resolutions through a multistage process of down-sampling the image to obtain various image resolutions [16].”
MethodLi, Y., Ma, C., Yan, Y., Zhu, W., & Yang, X. (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/cvpr52729.2023.01222 · Cited by 39
“For 3D GAN Inversion, we adopt the same metrics as 2D GAN inversion: L2 distance (or MSE loss) to calculate the pixel-wise similarity, LPIPS [58] distance to evaluate the perceptual similarity and MS-SSIM [53] to show the structural similarity.”
MethodContext-Based Trit-Plane Coding for Progressive Image Compression
Jeon, S., Choi, K. P., Park, Y., & Kim, C. (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/cvpr52729.2023.01379 · Cited by 45
“For MS-SSIM, we present decibel scores by MS-SSIM (dB) = − 10 · log 10 (1 − MS-SSIM ) .”
MethodInfluentialPerceptually Weighted Rate Distortion Optimization for Video-Based Point Cloud Compression
Zhang, Y., Ding, K., Li, N., et al. (2023) · IEEE Transactions on Image Processing · doi:10.1109/tip.2023.3327003 · Cited by 20
“(1) for channel φ , D from fine to coarse are set as 0.1, 0.6, and 0.3, respectively, according to [41].”
BackgroundInfluentialDehazed Image Quality Evaluation: From Partial Discrepancy to Blind Perception
Zhou, W., Zhang, R., Li, L., Yue, G., Gong, J., Chen, H., & Liu, H. (2024) · IEEE Transactions on Intelligent Vehicles · doi:10.1109/tiv.2024.3356055 · Cited by 24
“Besides, several variants of similarity measurement were also developed, such as the multiscale SSIM (MS-SSIM) index [27], the information content weighted SSIM (IW-SSIM) measure [28], the feature similarity (FSIM) index [29], the gradient similarity (GSM) model [30], the gradient magnitude…”
MethodA Reduced-Reference Quality Assessment Metric for Textured Mesh Digital Humans
Zhang, Z., Zhou, Y., Li, C., Fu, K., Sun, W., Liu, X., Min, X., & Zhai, G. (2024) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp48485.2024.10447636 · Cited by 23
“The former includes: PSNR, SSIM [21], MS-SSIM [22], GMSD [23], and G-LPIPS [3].”
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.