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 1,601–1,700 of 6,047
Huang, L., Song, C., Jin, G., et al. (2024) · IEEE Transactions on Aerospace and Electronic Systems · doi:10.1109/taes.2024.3380583 · Cited by 21
RankDVQA: Deep VQA based on Ranking-inspired Hybrid Training
Feng, C., Danier, D., Zhang, F., & Bull, D. (2024) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv57701.2024.00167 · Cited by 28
Study on subjective quality assessment of Screen Content Images
Shi, S., Zhang, X., Wang, S., et al. (2015) · Picture Coding Symposium · doi:10.1109/pcs.2015.7170050 · Cited by 20
“We apply the following 8 objective IQA metrics on the SCD: PSNR, SSIM [4], MS-SSIM [5], IW-SSIM [6], FSIM [9], VIF [16], GSM [17] and VSI [18] to investigate efficiency (a) (b)”
MethodInfluentialBinocular energy response based quality assessment of stereoscopic images
Shao, F., Jiang, G., Yu, M., Li, F., Peng, Z., & Fu, R. (2014) · Digital Signal Processing · doi:10.1016/j.dsp.2014.03.003 · Cited by 20
A Novel Vision-Based Adaptive Scanning for the Compression of Remote Sensing Images
Shi, C., Zhang, J., & Zhang, Y. (2015) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2015.2478145 · Cited by 21
“Thus, we adopt the MS-SSIM as an evaluation index in this paper.”
MethodInfluentialLim, C.-L., Paramesran, R., Jassim, W. A., Yu, Y.-P., & Ngan, K. N. (2016) · Journal of the Franklin Institute · doi:10.1016/j.jfranklin.2016.08.012 · Cited by 20
“In addition to that, we also tabulate the performance of five FR metrics: peaksignal-to-noise ratio (PSNR), structural similarity index (SSIM) [42], multi-scale structural similarity index (MS-SSIM) [44] and low-level feature similarity index (FSIM and FSIMc) [51].”
BackgroundA weighted full-reference image quality assessment based on visual saliency
Wen, Y., Li, Y., Zhang, X., Shi, W., Wang, L., & Chen, J. (2016) · Journal of Visual Communication and Image Representation · doi:10.1016/j.jvcir.2016.12.005 · Cited by 24
“Table 2: Performance comparison of compared IQA metrics on TID2008 database Method SROCC PLCC PSNR RMSE PSNR 0.5245 0.5309 47.0144 1.1372 VSNR 0.7046 0.6820 45.9660 0.9815 IFC 0.5692 0.7359 47.3207 0.9086 SSIM 0.7749 0.7732 50.1553 0.8511 MSSIM 0.8528 08425 50.9494 0.7299 VIF 0.7496 0.8090 50.1919 0.7888 FSIM 0.8805 0…”
BackgroundMethodInfluentialTraining-Free, Single-Image Super-Resolution Using a Dynamic Convolutional Network
Bhowmik, A., Shit, S., & Seelamantula, C. S. (2017) · IEEE Signal Processing Letters · doi:10.1109/lsp.2017.2752806 · Cited by 19
“Table I gives the PSNR, SSIM [35], multiscale SSIM (ms-SSIM) [36], and the morphological wavelet-based PSNR (mw-PSNR) [37] values averaged over the two datasets.”
MethodFull-Reference 3-D Video Quality Assessment Using Scene Component Statistical Dependencies
Appina, B., & Channappayya, S. S. (2018) · IEEE Signal Processing Letters · doi:10.1109/lsp.2018.2829107 · Cited by 20
A Study of Subjective Video Quality at Various Spatial Resolutions
Mackin, A., Afonso, M., Zhang, F., & Bull, D. (2018) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2018.8451225 · Cited by 22
“The image quality metrics considered were: MS-SSIM [26], PSNR [27], SSIM [28], VIF [29] and VSNR [30].”
MethodFQI: feature‐based reduced‐reference image quality assessment method for screen content images
Rahul, K., & Tiwari, A. K. (2019) · IET Image Processing · doi:10.1049/iet-ipr.2018.5496 · Cited by 20
“The most popular FR-IQA techniques for natural images are visual signal-to-noise ratio (VSNR) [9], local-tunedglobal model [10], structural similarity (SSIM) index [8, 11], multiscale-SSIM index [12], gradient-similarity [13], weightedSSIM [14], biologically inspired model [15], and feature-similarity index [16].”
MethodSISRSet: Single image super-resolution subjective evaluation test and objective quality assessment
Shi, G., Wan, W., Wu, J., Xie, X., Dong, W., & Wu, H. R. (2019) · Neurocomputing · doi:10.1016/j.neucom.2019.06.027 · Cited by 25
Powering Virtual Try-On via Auxiliary Human Segmentation Learning
Ayush, K., Jandial, S., Chopra, A., et al. (2019) · doi:10.1109/iccvw.2019.00397 · Cited by 19
Beyond Coding: Detection-driven Image Compression with Semantically Structured Bit-stream
He, T., Sun, S., Guo, Z., et al. (2019) · Picture Coding Symposium · doi:10.1109/pcs48520.2019.8954525 · Cited by 20
“We report Multi-Scale Structural Similarity (MS-SSIM) [7] between the original images and the reconstructed one to assess the performance of our model.”
BackgroundMethodVideo Content Analysis for Compliance Audit in Finance and Security Industry
Liu, B., Wu, M., Tao, M., Wang, Q., He, L., Shen, G., Chen, K., & Yan, J. (2020) · IEEE Access · doi:10.1109/access.2020.3005825 · Cited by 21
“Here in our system, the Mean Structural Similarity Index (MSSIM) [18] is used for measuring the difference between the two images, which is often more effective to capture the local changes of the image.”
BackgroundMethodSubjective Test Dataset and Meta-data-based Models for 360° Streaming Video Quality
Fremerey, S., Göring, S., Rao, R. R. R., et al. (2020) · IEEE International Workshop on Multimedia Signal Processing (MMSP) · doi:10.1109/mmsp48831.2020.9287065 · Cited by 20
“We further applied several other metrics, i.e. PSNR, SSIM [30, 31] and VIF [27].”
MethodRecent developments in computational color image denoising with PDEs to deep learning: a review
Salamat, N., Missen, M. M. S., & Prasath, V. B. S. (2021) · Artificial Intelligence Review · doi:10.1007/s10462-021-09977-z · Cited by 21
“Wang et al. (2017b) extended the combination of MAD and energy functional”
BackgroundDeep Perceptual Image Quality Assessment for Compression
Mier, J. C., Huang, E., Talebi, H., Yang, F., & Milanfar, P. (2021) · Proceedings - International Conference on Image Processing · doi:10.1109/icip42928.2021.9506217 · Cited by 22
“MSE, PSNR, SSIM, MS-SSIM [5] etc., are the dominant metrics used for measuring distortion resulting from compression.”
BackgroundInfluentialCombined Full-Reference Image Quality Metrics for Objective Assessment of Multiply Distorted Images
Okarma, K., Lech, P., & Лукин, В. В. (2021) · Electronics · doi:10.3390/electronics10182256 · Cited by 22
“…of weighting exponents for the product of three metrics [5] using the TID2008 database, although during further experiments, one of the metrics was replaced by FSIM forming the Combined Image Similarity Index (CISI) [6], being the weighted product of MS-SSIM [17], VIF [18] and FSIM [20].”
BackgroundMethodBenchmarking of objective quality metrics for point cloud compression
Lazzarotto, D. N., Alexiou, E., & Ebrahimi, T. (2021) · IEEE International Workshop on Multimedia Signal Processing (MMSP) · doi:10.1109/mmsp53017.2021.9733538 · Cited by 21
“From the latter category, we recruit the PSNR, SSIM [28], MS-SSIM [29], and VIFp [30] (i.e., multi-scale in pixel domain).”
MethodRealization of a Real-Time Image Denoising System for Dashboard Camera Applications
Yu, C., & Hou, L.-Z. (2022) · IEEE Transactions on Consumer Electronics · doi:10.1109/tce.2022.3175796 · Cited by 20
A New Method for Impulse Noise Elimination and Edge Preservation
Ramadan, Z. M. (2014) · Canadian Journal of Electrical and Computer Engineering · doi:10.1109/cjece.2014.2309071 · Cited by 22
Image quality assessment by preprocessing and full reference model combination
Bianco, S., Ciocca, G., Marini, F., & Schettini, R. (2009) · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · doi:10.1117/12.806693 · Cited by 22
An Overview of Perceptual Processing for Digital Pictures
Wu, H. R., Lin, W., & Karam, L. J. (2012) · IEEE International Conference on Multimedia and Expo workshops · doi:10.1109/icmew.2012.27 · Cited by 25
Wide-Band Color Imagery Restoration for RGB-NIR Single Sensor Images
Poma, X. S., Sappa, A. D., & Hammoud, R. I. (2018) · Sensors · doi:10.3390/s18072059 · Cited by 20
“The AE, as for other similarity measures [42,43], is computed over every single pixel of a pair of image improving MSE values; Ŷ is the RGB image predicted by the proposed networks, and Y is the corresponding ground truth image.”
BackgroundEdge Detection Techniques for Quantifying Spatial Imaging System Performance and Image Quality
van Zwanenberg, O., Triantaphillidou, S., Jenkin, R., et al. (2019) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw.2019.00238 · Cited by 20
From Patches to Pictures (PaQ-2-PiQ): Mapping the Perceptual Space of Picture QualityPreprint
Ying, Z., Niu, H., Gupta, P., Mahajan, D., Ghadiyaram, D., & Bovik, A. (2019) · Computer Vision and Pattern Recognition · doi:10.48550/arxiv.1912.10088 · Cited by 492
“Full-Reference models: Many full-reference (FR) perceptual picture quality predictors, which make comparisons against high-quality reference pictures, are available [9, 10], [27, 28, 29, 30, 31, 32, 33].”
BackgroundSelf-citationGeneralized Octave Convolutions for Learned Multi-Frequency Image CompressionPreprint
Akbari, M. A., Liang, J., Han, J., & Tu, C. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2002.10032 · Cited by 30
“We use both PSNR and MS-SSIMdB as the evaluation metrics, where MS-SSIMdB represents MS-SSIM scores in dB defined as: MS-SSIMdB = −10log10(1 −MS-SSIM).”
MethodInfluentialUniversal Rate-Distortion-Perception Representations for Lossy CompressionPreprint
Zhang, G., Qian, J., Chen, J., et al. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2106.10311 · Cited by 20
“Common full-reference metrics include MSE, SSIM/MS-SSIM [36, 37], PSNR or deep feature based distances [16, 41].”
Wang, C., Shang, K., Zhang, H., Li, Q., Hui, Y., & Zhou, S. K. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2111.10790 · Cited by 44
“5) [32,33] and Peak Signalto-Noise Ratio (PSNR).”
Virtual staining for pathology: Challenges, limitations and perspectives
Lin, W., Hu, Y., Zhu, R., Wang, B., & Wang, L. (2025) · Intelligent Oncology · doi:10.1016/j.intonc.2025.03.005 · Cited by 24
Wang, N., Geng, B., Shi, Y., Niu, Y., & Luo, A. (2025) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2025.3557939 · Cited by 19
“The multiscale structural similarity (SSIM) index (MS_SSIM) [51] is an extension of SSIM, which is a metric used to compare the SSIM of two pictures.”
Zhao, T., Zeng, K., Rehman, A., & Wang, Z. (2013) · Asilomar Conference on Signals, Systems and Computers (ACSSC) · doi:10.1109/acssc.2013.6810465 · Cited by 23
Self-citationBoosting dark channel dehazing via weighted local constant assumption
Zhu, M., He, B., Liu, J., & Yu, J. (2020) · Signal Processing · doi:10.1016/j.sigpro.2019.107453 · Cited by 20
Light Field Image Quality Assessment: An Overview
Huang, H., Zeng, H., Tian, Y., Chen, J., Zhu, J., & Ma, K. (2020) · IEEE International Conference on Multimedia Information Processing and Retrieval (MIPR) · doi:10.1109/mipr49039.2020.00077 · Cited by 21
“On this basis, Wang et al. [8] proposed the multi-scale structural similarity (MS-SSIM), and achieved better performance than the SSIM.”
MethodPerceptual Video Quality Prediction Emphasizing Chroma Distortions
Chen, L.-H., Bampis, C. G., Li, Z., Sole, J., & Bovik, A. C. (2020) · IEEE Transactions on Image Processing · doi:10.1109/tip.2020.3043127 · Cited by 29
Self-citationA Review on Application of Fourier Transform in Image Restoration
John, A. M., Khanna, K., Prasad, R. R., & Pillai, L. G. (2020) · International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (I-SMAC) · doi:10.1109/i-smac49090.2020.9243510 · Cited by 23
“The v isual quality of the image depends upon three key quality parameters of the image such as correlation, luminance and contrast [33][34][41].”
BackgroundConditional Generative ConvNets for Exemplar-Based Texture Synthesis
Wang, Z.-M., Li, M., & Xia, G.-S. (2021) · IEEE Transactions on Image Processing · doi:10.1109/tip.2021.3052075 · Cited by 24
“For quantitative evaluation, we measure multi-scale structural similarity [47] (MS-SSIM) between the synthesized tex-”
MethodSelf-citationWei, J., Li, M., Hu, Y., et al. (2021) · IEEE Geoscience and Remote Sensing Letters · doi:10.1109/lgrs.2020.3047789 · Cited by 19
Images denoising for COVID-19 chest X-ray based on multi-resolution parallel residual CNN
Jiang, X., Zhu, Y. M., Zheng, B., & Yang, D. (2021) · Machine Vision and Applications · doi:10.1007/s00138-021-01224-3 · Cited by 22
“We propose the MSL1 loss to train our MPR-CNN by adding multi-scale structural similarity (MS_SSIM) [62, 63] and L1 loss.”
MethodMikhailiuk, A., Ye, N., & Mantiuk, R. (2021) · Electronic Imaging · doi:10.2352/issn.2470-1173.2021.11.hvei-152 · Cited by 22
“We evaluated both hand-crafted image quality metrics (PSNR, SSIM [30], MS-SSIM [29], FSIM [36], HDRVQM [22]) and one metric based on machine learning (deep photometric image quality metric: PU-PieApp) [20].”
MethodMW-GAN+ for Perceptual Quality Enhancement on Compressed Video
Wang, J., Xu, M., Deng, X., Shen, L., & Song, Y. (2021) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2021.3128275 · Cited by 27
“There are many widely used full-reference distortion metrics, including PSNR, structural similarity index (SSIM) [28], multi-scale structural similarity (MS-SSIM) [29] and visual information fidelity (VIF) [30].”
Method(2022) · doi:10.3390/electronics11040559 · Cited by 19
“, 2stepQA [58], CSV [59], DISTS [60], GSM [8], MAD [57], MS-SSIM [16], ReSIFT [61], RVSIM [62], and SSIM [15], whose source codes were made online available for the research community.”
BackgroundMethodResearch on High-Resolution Face Image Inpainting Method Based on StyleGAN
He, L., Qiang, Z., Shao, X., Lin, H., Wang, M., & Dai, F. (2022) · Electronics · doi:10.3390/electronics11101620 · Cited by 19
Xian, W., Zhou, M., Fang, B., & Kwong, S. (2022) · Information Sciences · doi:10.1016/j.ins.2022.07.053 · Cited by 20
Saad, M. M., Rehmani, M. H., & O’Reilly, R. (2022) · Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) · doi:10.1109/embc48229.2022.9871260 · Cited by 24
“The α , β , and γ are the weight parameters as reported in [14].”
Moniruzzaman, M., Rassau, A., Chai, D., & Islam, S. M. S. (2022) · Journal of Intelligent & Robotic Systems · doi:10.1007/s10846-022-01749-3 · Cited by 21
“The literature [43, 45–47] suggests MS-SSIM is more robust and performs better for both images and video data.”
BackgroundMethodAn improved generative adversarial network for remote sensing image super‐resolution
Guo, J., Lv, F., Shen, J., et al. (2023) · IET Image Processing · doi:10.1049/ipr2.12760 · Cited by 19
FP-Conv-CM: Fuzzy Probabilistic Convolution C-Means
Moutaouakil, K. E., Palade, V., Safouan, S., & Charroud, A. (2023) · Mathematics · doi:10.3390/math11081931 · Cited by 20
Study of Subjective and Objective Quality Assessment of Mobile Cloud Gaming Videos
Saha, A., Chen, Y.-C., Davis, C., Qiu, B., Wang, X., Gowda, R., Katsavounidis, I. X., & Bovik, A. C. (2023) · IEEE Transactions on Image Processing · doi:10.1109/tip.2023.3281170 · Cited by 21
“Here, we comprehensively compare the performances of seven FR-VQA algorithms: PSNR, SSIM [46], MS-SSIM [47], ST-RRED [48], SpEED-QA [49], ST-GREED [50], and VMAF on the LIVEMeta Mobile Cloud Gaming database to explore for their”
MethodInfluentialSelf-citationContext and detail interaction network for stereo rain streak and raindrop removal
Nie, J., Xie, J., Cao, J., & Pang, Y. (2023) · Neural Networks · doi:10.1016/j.neunet.2023.07.013 · Cited by 18
Guo, Y., Liu, R. W., Lu, Y., Nie, J., Lyu, L., Xiong, Z., Kang, J., Yu, H., & Niyato, D. T. (2023) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2023.3298041 · Cited by 22
“More details on the definition of MS-SSIM can be found in [33].”
BackgroundMethodPractical Deep Dispersed Watermarking with Synchronization and Fusion
Guo, H., Zhang, Q., Luo, J.-W., Guo, F., Zhang, W., Su, X., & Li, M. (2023) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3581783.3612015 · Cited by 47
Perception-Driven Similarity-Clarity Tradeoff for Image Super-Resolution Quality Assessment
Zhang, K., Zhao, T., Chen, W., Niu, Y., Hu, J., & Lin, W. (2023) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2023.3341626 · Cited by 21
“The comparison methods consist of five classic generic full-reference IQA methods (SSIM [37], MS-SSIM [30], FSIM [38], GSM [39], GMSD [40]), five generic no-reference IQA methods (DIIVINE [41], BRISQUE [42], BLIINDS-II [43], MANIQA [44], FPR [45]), and six popular SR-IQA meth-ods (LNQM [5], HYQM…”
MethodPerceptual Quality Assessment of Face Video Compression: A Benchmark and An Effective Method
Li, Y., Chen, B., Chen, B., Wang, M., Wang, S., & Lin, W. (2024) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2024.3380260 · Cited by 30
“In specific, five FR-IQA algorithms, including PSNR [40], SSIM [19], Multi-scale Structural Similarity (MS-SSIM) [42], DISTS [26], and LPIPS [25] are selected, among which DISTS and LPIPS are two data-driven algorithms based on deep neural networks.”
MethodXSIM : A structural similarity index measure optimized for MRI QSM
Milovic, C., Cristián Tejos, N., Silva, J., Shmueli, K., & Pablo Irarrázaval, M. (2024) · Magnetic Resonance in Medicine · doi:10.1002/mrm.30271 · Cited by 23
NeRF View Synthesis: Subjective Quality Assessment and Objective Metrics Evaluation
Martin, P., Rodrigues, A., Ascenso, J. M., & Queluz, T. R. M. P. (2024) · IEEE Access · doi:10.1109/access.2024.3522768 · Cited by 22
“…was based on several criteria, notably: i) metrics that are typically applied in the NVS literature, namely PSNR-Y, PSNR-YUV, SSIM [5], MS-SSIM [6], LPIPS [7], and FovVideoVDP [8], since it is important to know whether they are in fact suitable for assessing the quality of NVS content; ii)…”
MethodSemantic-Importance-Aware Reordering-Enhanced Semantic Communication System With OFDM Transmission
Liu, Y., Dong, C., Liang, H., Li, W., Bao, Z., Zheng, Z., Xu, X., & Zhang, P. (2025) · IEEE Internet of Things Journal · doi:10.1109/jiot.2025.3529492 · Cited by 24
“2) Evaluation Metrics: PSNR [36] and MS-SSIM [37] were used to evaluate the quality of the reconstructed data.”
Validation Of A New Full Reference Metric For Quality Assessment Of Mobile 3Dtv Content
Boev, A., Egiazarian, K., Gotchev, A. P., & Jin, L. (2011) · Zenodo (CERN European Organization for Nuclear Research) · doi:10.5281/zenodo.42622 · Cited by 20
Multi-modal Color Medical Image Fusion Using Quaternion Discrete Fourier Transform
Nawaz, Q., Xiao, B., Hamid, I., et al. (2016) · Sensing and Imaging · doi:10.1007/s11220-016-0131-x · Cited by 19
“…with widely used existing image fusion algorithms by applying objective image quality assessment matrices [25] that includes Multiscale Structural Similarity (MSSSIM) [26], Peak Signal-to-Noise Ratio (PSNR) [27], Root-Mean-Square-Error (RMSE) [25], Mutual Information (MI) [28] and Entropy [29].”
MethodAtmospheric turbulence mitigation based on turbulence extraction
He, R., Wang, Z., Fan, Y., et al. (2016) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp.2016.7471915 · Cited by 19
“In addition, three full reference image quality metrics, Peak signal to noise ratio (PSNR), perception-based image model (PIM) [21] and multi-scale structural similarity (MSSSIM) [22], are utilized to quantitatively evaluate the performance of the comparison.”
MethodSelf-citation3D-HEVC visual quality assessment: Database and bitstream model
Zhou, W., Liao, N., Chen, Z., et al. (2016) · International Conference on Quality of Multimedia Experience (QoMEX) · doi:10.1109/qomex.2016.7498946 · Cited by 19
“Hence, we compared the proposed bitstream-layer No-Reference model with some widely used Full-Reference 2D quality assessment metrics, such as PSNR, FSIM [19], MS-SSIM [20] and VQM (i.”
MethodDCT-QM: A DCT-Based Quality Degradation Metric for Image Quality Optimization Problems
Bae, S., & Kim, M. (2016) · IEEE Transactions on Image Processing · doi:10.1109/tip.2016.2598492 · Cited by 22
“This framework was firstly devised by Wang and Bovik [7], [8] and has become a standard framework of many FR-IQA methods [9]–[16].”
MethodNo reference image quality assessment
Mandgaonkar, V. S., & Kulkarni, C. V. (2014) · IEEE India Conference · doi:10.1109/indicon.2014.7030678 · Cited by 19
Ding, Y., Zhao, Y., & Zhao, X. (2017) · Signal processing. Image communication · doi:10.1016/j.image.2017.03.001 · Cited by 17
Blind Quality Estimation by Disentangling Perceptual and Noisy Features in High Dynamic Range Images
Kottayil, N. K., Valenzise, G., Dufaux, F., & Cheng, I. (2017) · IEEE Transactions on Image Processing · doi:10.1109/tip.2017.2778570 · Cited by 21
“The Tone-Mapped Quality Index (TMQI) metric [13] follows the structural fidelity criterion [14], to compare an HDR image with its tone-mapped version, by embedding the knowledge of the Contrast Sensitivity Function (CSF) at different values of luminance [15], [16].”
MethodPrediction of Compression Ratio for DCT-Based Coders With Application to Remote Sensing Images
Zemliachenko, A., Kozhemiakin, R. A., Abramov, S., et al. (2017) · IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing · doi:10.1109/jstars.2017.2781906 · Cited by 19
New feature selection algorithms for no-reference image quality assessment
Nizami, I. F., Majid, M., & Khurshid, K. (2018) · Applied Intelligence · doi:10.1007/s10489-018-1151-0 · Cited by 19
“SSIM is extended to multi-scale and is named as MS-SSIM [14].”
MethodHybrid Intraprediction Based on Local and Nonlocal Correlations
Zhang, T., Fan, X., Zhao, D., et al. (2017) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2017.2775223 · Cited by 19
Reliability of the Most Common Objective Metrics for Light Field Quality Assessment
Amirpour, H., Pinheiro, A., Pereira, M., et al. (2019) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp.2019.8683215 · Cited by 21
“Six full reference quality metrics were computed to assess the performance of the state-of-the-art encoders including PSNR Y, PSNR YUV, SSIM Y [12], MS-SSIM Y [13], FSIM Y [14], and HDRVDP-2 Y [15].”
MethodDiscriminability Tests for Visualization Effectiveness and Scalability
Veras, R., & Collins, C. M. (2019) · IEEE Transactions on Visualization and Computer Graphics · doi:10.1109/tvcg.2019.2934432 · Cited by 22
“[42] found optimal MS-SSIM weight parameters for natural images (0.”
BackgroundInfluentialScale-aware Conditional Generative Adversarial Network for Image Dehazing
Kumar, P., Priyankar, S., & Sur, J. A. (2020) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv45572.2020.9093528 · Cited by 19
“…index (UQI) [46], Learned perceptual image patch similarity (LPIPS) [59], Mean squared error (MSE), Multi-scale structural similarity (MS-SSIM) [48], Feature similarity (FSIM) index [58], Color difference (CIEDE 2000) [42], Haar wavelet-based perceptual similarity index (HaarPSI) [36],…”
MethodÜstün, D. (2020) · Engineering Computations · doi:10.1108/ec-02-2020-0126 · Cited by 19
Accelerating Super-Resolution and Visual Task Analysis in Medical Images
Zamzmi, G., Rajaraman, S., & Antani, S. (2020) · Applied Sciences · doi:10.3390/app10124282 · Cited by 18
“Various studies [35,36] found that SSIM provides a better indication of SISR performance than PSNR because it can better represent human visual perception.”
BackgroundMethodBhat, N., Saggu, N., Pragati, P., & Kumar, S. (2020) · International Conference on Communication and Electronics Systems (ICCES) · doi:10.1109/icces48766.2020.9137895 · Cited by 19
“The metrics used will also include the structure similarity index (SSIM) [14], multiscale structure similarity index (MSSIM) [13] and peak signal-to-noise ratio (PSNR).”
MethodMTF-CRNN: Multiscale Time-Frequency Convolutional Recurrent Neural Network for Sound Event Detection
Zhang, K., Cai, Y., Ren, Y., Ye, R., & He, L. J. (2020) · IEEE Access · doi:10.1109/access.2020.3015047 · Cited by 20
“The multiscale filter is a time-frequency local transform that can extract information of features from signals effectively and efficiently [26]–[28].”
BackgroundDual In-painting Model for Unsupervised Gaze Correction and Animation in the Wild
Zhang, J., Chen, J., Tang, H., Wang, W., Yan, Y., Sangineto, E., & Sebe, N. (2020) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3394171.3413981 · Cited by 21
“GazeGAN and PRGAN obtain the best results with 1.0 for MSSSIM and 0.0 for LPIPS.”
MethodInfluentialControllable Multi-Attribute Editing of High-Resolution Face Images
Deng, Q., Li, Q., Cao, J., Liu, Y., & Sun, Z. (2020) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2020.3033184 · Cited by 19
“We compare quantitative results both at 128 and 512 resolution in terms of Fréchet Inception Distance (FID) [42] and Multi-Scale Structural SIMilarity (MS-SSIM) [43].”
BackgroundMethodYang, J., Bian, Z., Zhao, Y. S., Lu, W. B., & Gao, X. (2021) · IEEE Transactions on Broadcasting · doi:10.1109/tbc.2021.3064266 · Cited by 21
Yudina, E. N., Petrovskaia, A., Shadrin, D., Tregubova, P., Chernova, E., Pukalchik, M., & Oseledets, I. V. (2021) · Water · doi:10.3390/w13070888 · Cited by 21
“The structural similarity index is calculated for various windows of an image [34].”
BackgroundChady, T., Okarma, K., Mikołajczyk, R., et al. (2021) · Materials · doi:10.3390/ma14164452 · Cited by 19
“Following the first success of the SSIM, many similar approaches and extensions of this metric were proposed by various researchers during recent years, starting from its multi-scale version, known as MS-SSIM [26], making its results independent of the sliding window size.”
MethodA Deeply Modulated Scheme for Variable-Rate Video Compression
Lin, J., Liu, D., Liang, J., et al. (2021) · Proceedings - International Conference on Image Processing · doi:10.1109/icip42928.2021.9506269 · Cited by 19
“Both PSNR and MS-SSIM [18] are used as the quality metrics.”
MethodMultitask Identity-Aware Image Steganography via Minimax Optimization
Cui, J., Zhang, P., Li, S., Zheng, L., Bao, C., Xia, J., & Li, X. (2021) · IEEE Transactions on Image Processing · doi:10.1109/tip.2021.3107999 · Cited by 21
iDAM: Iteratively Trained Deep In-loop Filter with Adaptive Model Selection
Li, Y., Zhang, L., & Zhang, K. (2022) · ACM Transactions on Multimedia Computing Communications and Applications · doi:10.1145/3529107 · Cited by 19
“Table 3 gives overall performances of the proposed techniques in terms of MS-SSIM and VMAF.”
MethodInfluentialImage visual quality metrics verification by TID2013: Exploring of mean square error drawbacks
Ponomarenko, N. N., Лукин, В. В., Єремеєв, О., Vozel, B., Chehdi, K., Egiazarian, K., & Astola, J. T. (2014) · European Workshop on Visual Information Processing · doi:10.1109/euvip.2014.7018376 · Cited by 19
Video Compression through Image InterpolationPreprint
Wu, C.-Y., Singhal, N., & Krähenbühl, P. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1804.06919 · Cited by 19
“Our deep video codec outperforms all deep learning baselines, MPEG-4 Part 2, and H.261 in both compression rate and visual quality measured by MS-SSIM [29] and PSNR.”
MethodDeep Perceptual CompressionPreprint
Patel, Y., Appalaraju, S., & Manmatha, R. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1907.08310 · Cited by 27
“In Mentzer et al [27], MSSSIM[45] is used for measuring distortion between images D. Checkerboard patterns: Image generation models with deconvolution based up-sampling are known to generate certain checkerboard patterns depending on the loss function.”
BackgroundMethodInfluentialHuman Perceptual Evaluations for Image CompressionPreprint
Patel, Y., Appalaraju, S., & Manmatha, R. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1908.04187 · Cited by 19
“More recently, the trend is to evaluate compression using multiscale structure similarity (MS-SSIM) [19] sample compression works [1, 2, 11, 13, 10].”
Perceptual Adversarial Robustness: Defense Against Unseen Threat ModelsPreprint
Laidlaw, C., Singla, S., & Feizi, S. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2006.12655 · Cited by 146
“However, these similarity measures disagree with human vision on many types of perturbations such as blurring and spatial transformations, which has motivated others such as SSIM [21], MS-SSIM [30], CW-SSIM [31], HDR-VDP-2 [32] and LPIPS [22].”
BackgroundLivada, Č., Horvat, T., & Baumgartner, A. (2023) · Applied Sciences · doi:10.3390/app13053152 · Cited by 20
Monitoring Outdoor Parking in Urban Areas With Unmanned Aerial Vehicles
Kim, S., Tak, Y., Barmpounakis, E., & Geroliminis, N. (2024) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2024.3397588 · Cited by 21
“To quantify the similarity between the crops c1 and c2, we rely on the structural similarity index measure (SSIM) [62], a widely used metric for assessing the similarity between two images.”
MethodsDAC—Semantic Digital Analog Converter for Semantic Communications
Bao, Z., Liang, H., Liu, X., Li, C., Dong, C., Xu, X., Guo, C., Chen, H., & Zhang, P. (2025) · IEEE Transactions on Communications · doi:10.1109/tcomm.2025.3566976 · Cited by 26
“This system is designed for image transmission, and the end-to-end communication performance can be measured using the peak signal-to-noise ratio (PSNR) and multi-scale structure similarity (MS-SSIM) [31].”
Tackling the Sheer Scale of Subjective QoE
Menkovski, V., Exarchakos, G., & Liotta, A. (2012) · Lecture notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering · doi:10.1007/978-3-642-30419-4_1 · Cited by 21
“Some effective FR methods include MultiScale-Structural SIMilarity index (MS-SSIM) [2], Perceptual Video Quality Metric (PVQM) [3], and the perceptual spatio-temporal frequency-domain based MOtion-based Video Integrity Evaluation (MOVIE) [4].”
MethodSaliency weighted quality assessment of tone-mapped images
Nasrinpour, H. R., & Bruce, N. D. B. (2015) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2015.7351748 · Cited by 20
Rate Distortion Characteristic Modeling for Neural Image Compression
Jia, C., Ge, Z., Wang, S., Ma, S., & Gao, W. (2022) · DCC · doi:10.1109/dcc52660.2022.00028 · Cited by 21
“It achieves more pleasant perceptual quality using MS-SSIM based compression.”
MethodInfluentialImmersive Video Coding: Should Geometry Information Be Transmitted as Depth Maps?
Garus, P., Henry, F., Jung, J., et al. (2021) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2021.3100006 · Cited by 18
“Beyond the CTC, MPEG-I Visual further studies the combination of additional metrics with the BD-Rate by replacing the PSNR with Video Multimethod Assessment Fusion (VMAF) [37] and Multi-Scale Structural Similarity (MS-SSIM) [38].”
Recent trending on learning based video compression: A survey
Hoang, T. M., & Zhou, J. (2021) · Cognitive Robotics · doi:10.1016/j.cogr.2021.08.003 · Cited by 25
UnfairGAN: An enhanced generative adversarial network for raindrop removal from a single image
Nguyen, D. M., Le, T. P., Vo, D. M., & Lee, S.-W. (2022) · Expert Systems with Applications · doi:10.1016/j.eswa.2022.118232 · Cited by 20
Medical image fusion based on DTNP systems and Laplacian pyramid
Mi, S., Zhang, L., Peng, H. J., & Wang, J. (2021) · Journal of Membrane Computing · doi:10.1007/s41965-021-00087-x · Cited by 18
“7 The results fused by the different methods for the 5th group of images used quantitative metrics are utilized to evaluate the proposed and compared fusion methods, including Qabf [37], Mutual Information (MI) [38], Feature Mutual Information (FMI) [39], Sum of Correlation Differences (SCD) [40], Multiscale-Structura…”
MethodSpatial resolution improved fluorescence lifetime imaging via deep learning
Xiao, D., Zang, Z., Xie, W., et al. (2022) · Optics Express · doi:10.1364/oe.451215 · Cited by 18
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.