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 4,401–4,500 of 6,047
Dual In-painting Model for Unsupervised Gaze Correction and Animation in the WildPreprint
Zhang, J., Chen, J., Tang, H., et al. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2008.03834 · Cited by 2
“GazeGAN and PRGAN obtain the best results with 1.0 for MSSSIM and 0.0 for LPIPS.”
MethodInfluentialGenerator Versus Segmentor: Pseudo-healthy SynthesisPreprint
Yunlong, Z., Chenxin, L., Xin, L., Liyan, S., Zhuang, Y., Yue, H., Xinghao, D., Liu, X., & Yizhou, Y. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2009.05722 · Cited by 3
“…is expressed as the structural similarity calculated on normal pixels, which is defined as iD = MS-SSIM[(1 − y t ) G ( x p ) , (1 − y t ) (cid:12) x p ], where the (cid:12) denotes the pixel-wise multiplication, and the MS-SSIM() representes a masked Multi-Scale Structural Similarity Index [14].”
Posterior Sampling for Image Restoration using Explicit Patch PriorsPreprint
Friedman, R., & Weiss, Y. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2104.09895 · Cited by 5
“In recent years, however, the limitations of using the peak signal-to-noise ratio (PSNR) or related evaluation scores such as the structural similarity index measure [14] (SSIM) have become increasingly clear.”
BackgroundRobust Representation Learning via Perceptual Similarity MetricsPreprint
Taghanaki, S. A., Choi, K., Khasahmadi, A. H., & Goyal, A. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2106.06620 · Cited by 26
“Specifically, we compute the MS-SSIM metric as: where c j and s j denote the contrast and structure of images x and y at scale j respectively, and l M denotes the luminance only at scale M .”
InfluentialLearning from Synthetic Data for Opinion-free Blind Image Quality Assessment in the WildPreprint
Wang, Z., Tang, Z., Zhang, J., & Fang, Y. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2106.14076 · Cited by 2
“Based on the accessibility of original reference images, existing IQA methods can be categorized into full-reference IQA (FR-IQA) [51], [52], [61], reduced-reference IQA (RR-IQA) [50] and no-reference/blind IQA (NR-IQA/BIQA) [23], [24], [25], [66].”
MethodSelf-citationPerceptual Video Compression with Recurrent Conditional GANPreprint
Yang, R., Gool, L. V., & Timofte, R. (2021) · arXiv (Cornell University) · Cited by 2
ER-IQA: Boosting Perceptual Quality Assessment Using External Reference ImagesPreprint
Guo, J., Wang, W., Yang, W., Liao, Q., & Zhou, J. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2105.02464 · Cited by 3
“We also compare the ER-IQA model against FR-IQA meth-ods: SSIM [32], MS-SSIM [47], VIF [33], GMSD [34], FSIMc [48], DeepQA [49], WaDIQaM-FR [43] and DRF-IQA [50].”
MethodAn evaluation of existing and emerging digital image and video quality metrics.
Dosselmann, R. (2006) · doi:10.82465/4751 · Cited by 2
Tian, Y., Liu, F., Pang, G., et al. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2109.01303 · Cited by 2
“…combines three loss functions: 1) two reconstruction losses based on local and global multi-scale structural similarity index measure (MS-SSIM) [66] and mean absolute error (MAE) to train the encoder f θ : X → Z and decoder g ϕ : Z → X , 2) a regularisation loss to train adversarial…”
BackgroundBokeh Effect Rendering with Vision TransformersPreprint
Nagasubramaniam, H., & Younes, R. (2022) · doi:10.36227/techrxiv.17714849.v1 · Cited by 2
Camera Quality Assessment in Real-World ConditionsPreprint
Tworski, M., Pochon, B., & Lathuilière, S. (2022) · SSRN Electronic Journal · doi:10.2139/ssrn.4166549 · Cited by 2
Investigating the Visual Utility of Differentially Private ScatterplotsPreprint
Panavas, L., Crnovrsanin, T., Ullman, J., Adams, J. L., Sarvghad, A., Tory, M. K., & Dunne, C. (2023) · doi:10.31219/osf.io/b5zvn · Cited by 2
CID22: Large-Scale Subjective Quality Assessment for High Fidelity Image CompressionPreprint
Sneyers, J., Baruch, E. B., & Vaxman, Y. (2023) · doi:10.36227/techrxiv.22659061.v1 · Cited by 2
Artifact generation when using perceptual loss for image deblurringPreprint
Krawczyk, P., Gaertner, M., Jansche, A., Bernthaler, T., & Schneider, G. (2023) · doi:10.36227/techrxiv.23791962.v1 · Cited by 2
License Plate Super-Resolution Using Diffusion ModelsPreprint
AlHalawani, S., Benjdira, B., Ammar, A., et al. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2309.12506 · Cited by 2
“In order to assess the quality of our model comprehensively, we used the following metrics: Peak Signal-to-Noise Ratio (PSNR), Structural Similarity (SSIM) [32] and Multi-scale Structural Similarity (MS-SSIM) [33].”
MethodInfluentialGummadi, D., Chan, P. H., Wang, H., et al. (2023) · doi:10.36227/techrxiv.24566371 · Cited by 5
Comprehensive Image Quality Assessment (IQA) of JPEG, WebP, HEIF and AVIF FormatsPreprint
Kunwar, S. (2024) · OSF Preprints (OSF Preprints) · doi:10.31219/osf.io/ud7w4 · Cited by 5
Multi-Scale Attention in Attention Neural Network for Single Image DeblurringPreprint
Lee, H. S., & Cho, S. I. (2024) · SSRN Electronic Journal · doi:10.2139/ssrn.4756113 · Cited by 2
Barba, T., Bagley, B. A., Steyaert, S., et al. (2025) · medRxiv · doi:10.1101/2025.02.24.25322787 · Cited by 2
A review of generative coverless image steganography based on diffusion modelsPreprint
Zhang, S., Xue, J., Fu, J., et al. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5570333 · Cited by 2
Quality of Experience for High Definition Presentations - Case: Digital Cinema
Perkis, A., Rahayu, F. N., Reiter, U., et al. (2010) · Signals and communication technology · doi:10.1007/978-3-642-12802-8_2 · Cited by 2
“Additionally, a multi-scale SSIM (MSSIM) measure has been proposed based on the SSIM algorithm [20].”
MethodClustering of Temporal and Visual Data: Recent Advancements
Mudgal, P. (2026) · Data · doi:10.3390/data11010007 · Cited by 2
Exploration-Driven Generative Interactive Environments
Savov, N., Kazemi, N., Mahdi, M., Paudel, D. P., Wang, X., & Van Gool, L. (2025) · 2025 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52734.2025.02570 · Cited by 2
Signature and Log-Signature for the Study of Empirical Distributions Generated with GANs
de Curtò, J., de Zarzà, I., Roig, G., & Calafate, C. T. (2023) · Electronics · doi:10.3390/electronics12102192 · Cited by 6
Research on Structural Similarity for Image Quality Assessment
Hui, D., Ji, X., & Song, Y. (2011) · Lecture notes in electrical engineering · doi:10.1007/978-1-4471-2386-6_212 · Cited by 1
Pang, K., Shi, Z., & Zhao, L. (2017) · International Conference on Information Science and Control Engineering · doi:10.1109/icisce.2017.46 · Cited by 1
“We also compare the performance of the proposed scheme with the state-of-the-art IQA schemes, including three state-of-the-arts (SSIM [6], MS-SSIM [7], and IW-SSIM [8]).”
BackgroundMethodKumar, V., Mukherjee, J., & Mandal, S. K. D. (2018) · Springer eBooks · doi:10.1007/978-981-10-7221-5_5 · Cited by 1
Full Reference Quality Assessment of Full HD Images Using Combined Saliency Priors in Multi-Scale
Khan, S. (2018) · National Conference on Communications · doi:10.1109/ncc.2018.8600203 · Cited by 1
“Traditional FRIQA metrics like the structural similarity (SSIM) index [5] and the multi-scale SSIM (MS-SSIM) index [6], proposed by Wang et al.”
MethodMultiresolution Mixture Generative Adversarial Network For Image Super-Resolution
Wang, Y., Lan, X., Zhang, Y., et al. (2020) · IEEE International Conference on Multimedia and Expo (ICME) · doi:10.1109/icme46284.2020.9102972 · Cited by 1
Small animal PET to CT image synthesis based on conditional generation network
Li, J., Wang, Y., Yang, Y., Zhang, X., Qu, Z., & Hu, S. (2021) · International Congress on Image and Signal Processing (CISP) · doi:10.1109/cisp-bmei53629.2021.9624232 · Cited by 4
Autoencoder-based bone removal algorithm from x-ray images of the lung
Kalisz, S., & Marczyk, M. (2021) · IEEE International Conference on Bioinformatics and Bioengineering (BIBE) · doi:10.1109/bibe52308.2021.9635451 · Cited by 3
“The formula proposed in [15] and used in [5], that minimizes the Mean Square Error (MSE) and maximizes the Structural Similarity Index Measure (SSIM), was used as a loss function.”
MethodStand-in Backdoor: A Stealthy and Powerful Backdoor Attack
Li, S., Li, H., & Chen, H. (2021) · Globecom · doi:10.1109/globecom46510.2021.9685762 · Cited by 1
“C 1 , C 2 and C 3 are constants, where the settings of them are included in [15].”
BackgroundJoint Contrast Enhancement and Noise Suppression of Low-light Images Via Deep Learning
Gao, Y., Zheng, Y., Su, J., & Bao, M. (2022) · International Conference on Robotics and Automation Sciences (ICRAS) · doi:10.1109/icras55217.2022.9842021 · Cited by 1
Lightweight image super-resolution with multiscale residual attention network
Xiao, C., Dong, H. F., Li, H., et al. (2022) · Journal of Electronic Imaging · doi:10.1117/1.jei.31.4.043028 · Cited by 1
“…time, we also adopt another four evaluation matrices, namely information fidelity criterion (IFC), 40 noise quality measure (NQM), 41 weighted peak signal-to-noise ratio (WPSNR), and multiscale structure similarity index (MSSSIM), 42 which obtain high correlation with the human perceptual score.”
MethodA Soft-ranked Index Fusion Framework with Saliency Weighting for Image Quality Assessment
Yu, L., Wang, Z., Yan, Y., Zhu, L., & Wang, S. (2022) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw56347.2022.00197 · Cited by 1
“Structural similarity (SSIM) [1] and its variants (e.g., MS-SSIM [2] and FSIM [3]) consider image degradation as a perceived change in structural information.”
BackgroundMethodSelf-citationA Two-Level Auto-Encoder for Distributed Stereo Coding
Harel, Y., & Avidan, S. (2022) · IEEE International Conference on Computational Photography (ICCP) · doi:10.1109/iccp54855.2022.9887724 · Cited by 1
“We too will use the MS-SSIM measure.”
BackgroundInfluentialDomain-Specific Fusion Of Objective Video Quality Metrics
Chadha, A., Katsavounidis, I. X., Bhunia, A. K., et al. (2022) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3503161.3548375 · Cited by 1
“Common metrics for distortion include PSNR, SSIM [30], MS-SSIM [31], information fidelity criterion (IFC) [22] and visual information fidelity (VIF) [21].”
MethodLocal and Global Fusion Network For Learned Image Compression
Zhang, G., Zhang, X., & Zhu, S. (2022) · Proceedings - International Conference on Image Processing · doi:10.1109/icip46576.2022.9897597 · Cited by 1
“In our experiments, we optimize our model based on two different image quality metrics, PSNR and MS-SSIM [22].”
MethodInfluentialDGS-Fuse: unsupervised image fusion network combining global information
Wu, K., Wang, M., Mei, Y., Shan, H., & Xu, Z. (2022) · Journal of Electronic Imaging · doi:10.1117/1.jei.31.6.063045 · Cited by 1
Screen content image quality assessment using curvelet transform
Loh, W.-T., & Bong, D. B. L. (2022) · Signal Image and Video Processing · doi:10.1007/s11760-022-02415-9 · Cited by 1
“For the NI-based QA methods, SSIM [19], MSSIM [40], IW-SSIM [41], FSIM [42], GMSD [43], and VIF [44] are used.”
MethodQuality assessment for microscopic parasite images
Asri, M. A. A., Mokhtar, N. B., Rajagopal, H., et al. (2023) · Proceedings of International Conference on Artificial Life and Robotics · doi:10.5954/icarob.2023.os1-4 · Cited by 1
Mori, S., Hirai, R., Sakata, Y., Koto, M., & Ishikawa, H. (2023) · Physical and Engineering Sciences in Medicine · doi:10.1007/s13246-023-01320-w · Cited by 1
“The SSIM value has been shown to reflect human visual perception well [ 35 ].”
BackgroundA kidney staining transfer model based on CycleGan using Spade as the normalization layer
Wang, N., & Yin, S. (2023) · International Conference on Frontiers of Electronics, Information and Computation Technologies · doi:10.1109/icfeict59519.2023.00045 · Cited by 1
“Then, to quantify the degree of structural deformation of the image organization area, this section uses the structural similarity index measure (SSIM) method [18].”
MethodA Boundary-Guided and Lightweight Object Detector for Sonar Images
Zhou, X., Wei, Y., & Ouyang, W. (2023) · MTS/IEEE OCEANS Conference (OCEANS) · doi:10.1109/oceanslimerick52467.2023.10244705 · Cited by 1
“combining binary cross entropy loss [30], structural similarity loss [31], and IoU loss [32] is designed to implicitly guide the network to focus more attention on the target boundary information so that BASNet can obtain fine disguised target boundaries without explicit boundary extraction.”
BackgroundDual-Channel Semi-Supervised Adversarial Network for Building Segmentation from UAV-Captured Images
Zhang, W., Wu, C., Man, W., & Liu, M. (2023) · Remote Sensing · doi:10.3390/rs15235608 · Cited by 1
A video codec based on background extraction and moving object detection
Hadi, S., Shahbahrami, A., & Azgomi, H. (2023) · Multimedia Tools and Applications · doi:10.1007/s11042-023-17933-y · Cited by 1
“The exponents M , j , and j are used to adjust the relative importance of different components [62, 63].”
Gupta, T. (2024) · International Journal of Advances in Electrical Engineering · doi:10.22271/27084574.2024.v5.i1a.47 · Cited by 1
Xuyuan, F. A. N. G., Wei, J., Fu, R., Li, G., Caifen, H. E., Caoqian, Y. I., & 宁波市镇海区气象局, 宁. 3. Z. D. M. B. (2023) · National Remote Sensing Bulletin · doi:10.11834/jrs.20221621 · Cited by 2
Liu, J., Zhang, R., Carass, A., et al. (2024) · SPIE Medical Imaging: Clinical and Biomedical Imaging · doi:10.1117/12.3008361 · Cited by 1
Using diffusion model to generate high-resolution MRI
Chang, C., Pan, S., Peng, J., Salari, E., Roper, J. R., Qiu, R. L. J., Gao, Y., Liu, T., Shu, H. G., Mao, H., et al. (2024) · SPIE Medical Imaging: Clinical and Biomedical Imaging · doi:10.1117/12.3006586 · Cited by 1
“This assessment encompassed key metrics such as mean absolute error (MAE) [29] , noise quality measure (NQM) [30] , peak signal-to-noise ratio (PSNR) [31] , and multi-scale structural similarity (MSSIM) [32] .”
BackgroundEnhancing image quality prediction with self‐supervised visual masking
Çoğalan, U., Bemana, M., Seidel, H., & Myszkowski, K. (2024) · Computer Graphics Forum · doi:10.1111/cgf.15051 · Cited by 1
“Note for MS-SSIM, we used the same F across all scales, while the inputs are images at different scales.”
MethodInfluentialPerception-Aware Texture Similarity Prediction
Wang, W., & Dong, X. (2024) · IEEE Transactions on Image Processing · doi:10.1109/tip.2024.3404854 · Cited by 1
“Traditionally, texture descriptors were normally designed based on the response of filters [4], [5], the statistics of gray level values [6], the structural characteristics [1], [7], [8] and a texture model [9].”
BackgroundMethodHu, X., Bai, J., Chen, C., & Yu, H.-Y. (2024) · PeerJ Computer Science · doi:10.7717/peerj-cs.2083 · Cited by 1
“In Table 2, it can be observed that in both databases, traditional methods like SSIM ( Wang et al., 2004 ) and MS-SSIM ( Wang, Simoncelli & Bovik, 2003 ) perform poorly compared to the MB-FR-SIQA method due to their insuf fi cient consideration of stereo characteristics.”
MethodA Method for In-Loop Video Coding Restoration
Salazar, C., Trujillo, M., & Branch, J. W. (2024) · Electronics · doi:10.3390/electronics13122422 · Cited by 1
65‐1: Invited Paper: The Advantages of “Clocked‐Analog” Video Transport in Source Driver ICs
Henzen, A., Rockoff, T. E., Clarke, R., et al. (2024) · SID Symposium Digest of Technical Papers · doi:10.1002/sdtp.17138 · Cited by 1
“…by an optical system that can do image fidelity analysis (i.e. determine the optical errors and the associated visibility by human test-persons) [2][3], and if needed followed by image analysis and correction (similar to de-mura measurements on modern OLED displays which is capable of measuring…”
MethodGupta, S., Mukhopadhyay, T., Varade, D. M., et al. (2024) · Springer proceedings in physics · doi:10.1007/978-981-97-4557-9_11 · Cited by 1
Near-Offset Gap Trace Extrapolation Based on Self-Supervised Learning
Park, J., Kim, S., Seol, S. J., & Byun, J. (2024) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2024.3426599 · Cited by 1
“To further enhance the model optimization, we introduce a multiscale structural similarity index (MS-SSIM) [47] loss function to assign higher weights to the trace reconstruction where the main events.”
MethodVideo compression and optimization technologies - review
Uddin, S., Leszczuk, M., & Grega, M. (2024) · International Journal of Electronics and Telecommunications · doi:10.24425/ijet.2024.149603 · Cited by 2
“There are established objective quality assessment approaches that are the Peak-to-Noise Ratio (PSNR), Structural similarity metric (SSIM) [21], Mul-ti-Scale Structural Similarity [22], SSIMplus [23], Video Quality Model (VQM) [21], and Natural Image Quality Evaluator (NIQE) [24].”
BackgroundA Novel Loss Function for Deep Learning-Based One-Step Phase Unwrapping
Ye, X., & Yu, H. (2024) · IEEE International Geoscience and Remote Sensing Symposium (IGARSS) · doi:10.1109/igarss53475.2024.10641071 · Cited by 1
SANERV: Scene-Adaptive Neural Representation for Videos
Rhee, H., Chung, H., Jo, J. H., et al. (2024) · Proceedings - International Conference on Image Processing · doi:10.1109/icip51287.2024.10647620 · Cited by 1
Asymmetric Multi-Scale Dilated Attention based Network for Remote Sensing Image Compression
Ma, H., Zhou, P., Jia, J., et al. (2024) · International Conference on Ubiquitous Communication (Ucom) · doi:10.1109/ucom62433.2024.10695911 · Cited by 1
“We employ PSNR [15], MS-SSIM [16], BD-rate [17], and BD-PSNR [18] as objective metrics to evaluate the quality of the compression and reconstruction of our models.”
MethodBRIC: Bottom-Up Residual Vector Quantization for Learned Image Compression
Flowers, B., & Dey, S. (2024) · IEEE Access · doi:10.1109/access.2024.3476462 · Cited by 1
“They were pre-trained on the Vimeo90k [59] dataset, and the models optimized for Multi-Scale SSIM [60] were utilized as it is most similar to the metrics used in this work (there are no available pre-trained models for the LPIPS objective).”
MethodHybrid Wavelet‐Deep Learning Framework for Fluorescence Microscopy Images Enhancement
Branciforti, F., Maggiore, M., Meiburger, K. M., Pannellini, T., & Salvi, M. (2024) · International Journal of Imaging Systems and Technology · doi:10.1002/ima.23212 · Cited by 2
Evaluating Image Synthesis: A Modest Review of Techniques and Metrics
de Sousa, R. N., & Oliveira, S. A. F. (2024) · SIBGRAPI Conference on Graphics, Patterns and Images (SIBGRAPI) · doi:10.5753/sibgrapi.est.2024.31649 · Cited by 1
High-Quality Short-Range Radar Imaging with Coprime Sampling
Wang, Y., Zeng, T., Zhan, X., et al. (2024) · Remote Sensing · doi:10.3390/rs16244657 · Cited by 1
Verma, K., Srivastava, S., & Mishra, R. K. (2024) · Lecture notes in electrical engineering · doi:10.1007/978-981-97-4359-9_71 · Cited by 1
VCNPU: An Algorithm-Hardware Co-Optimized Framework for Accelerating Neural Video Compression
Zhang, S., Mao, W., & Wang, Z. (2024) · IEEE Transactions on Very Large Scale Integration (VLSI) Systems · doi:10.1109/tvlsi.2024.3515113 · Cited by 1
“We use bit per pixel (bpp) to measure the average number of bits used for motion and residual coding per pixel in each frame, and PSNR and MS-SSIM [39] to evaluate the distortion between the reconstructed frame and ground-truth.”
Self-citationCui, B., Zhang, Y., Yang, Q., & Xu, Y. (2024) · ACM International Conference on Multimedia in Asia (MMAsia) · doi:10.1145/3696409.3700162 · Cited by 2
“For instance, multi-scale structural similarity (MS-SSIM) [34] and multi-scale graph similarity (MS-GraphSIM) [46] have been developed to quantify image and point cloud distortions from the multi-scale perspective and achieved more robust performance than original single-scale versions.”
Lightweight Efficient Rate-Adaptive Network for Compression-Aware Image Rescaling
Li, D., Zhang, Y., & Liu, Y. (2025) · IEEE Signal Processing Letters · doi:10.1109/lsp.2025.3530853 · Cited by 1
Learned Image Compression With Efficient Cross-Platform Entropy Coding
Yang, R., Liu, D., Wu, F. H., et al. (2025) · IEEE Journal on Emerging and Selected Topics in Circuits and Systems · doi:10.1109/jetcas.2025.3538652 · Cited by 1
“Some very recent works [11], [12], [13], [14] can even surpass VVC [17], which is one of the best handcrafted image and video coding standards at present [18], on both Peak Signal-to-Noise Ratio (PSNR) and Multi-Scale Structural Similarity (MS-SSIM) [19].”
Day, J. D., Su, W., Shragge, J., et al. (2025) · AIAA SciTech Forum · doi:10.2514/6.2025-1220 · Cited by 1
SOFNet: A Salient Object-Enhanced Image Fusion Network
Chen, J., Xiong, A., Huang, W., & Zheng, W. (2024) · International Conference on Electronic Information Engineering and Computer Technology (EIECT) · doi:10.1109/eiect64462.2024.10866062 · Cited by 1
“Contemporary state-of-the-art metrics for assessing image structural similarity encompass the structural similarity index (SSIM) [12] and the multi-scale structural similarity index (MS-SSIM) [13].”
CSMR: A Multi-Modal Registered Dataset for Complex Scenarios
Li, C. J., Gao, K., Hu, Z., Yang, Z., Cai, M., Cheng, H., & Zhu, Z. (2025) · Remote Sensing · doi:10.3390/rs17050844 · Cited by 1
Chen, Z., Zhang, Z., Zuo, D., Wang, N., Fan, L., & Liu, Z. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10889254 · Cited by 1
“We compare these fused images with the standard fused images and calculate metrics such as mean square error (MSE) [27], structural similarity index measure (SSIM) [28], and normalized cross-correlation (NCC) [29].”
An adaptive finite difference method for total variation minimization
Jacumin, T., & Langer, A. (2025) · Numerical Algorithms · doi:10.1007/s11075-025-02044-6 · Cited by 1
“The measure between two windows x and y of common size ω × ω is [43]: , where µ x , µ y is the pixel sample mean of x, y respectively, σ 2 x , σ 2 y is the variance of x, y respectively, σ xy is the covariance of x and y , c 1 = 0 .”
BackgroundRamachandran, P., Anderson, D., Colbert, Z. M., Arrington, D., Huo, M., Pinkham, M. B., Foote, M. C., & Fielding, A. L. (2025) · Journal of Medical Physics · doi:10.4103/jmp.jmp_140_24 · Cited by 2
“[32] where μ x and μ y are the average intensity values of x and y, σ x 2 and σ y 2 are their variances, and σ xy is the covariance of images x and y, C 1 and C 2 are constants to stabilize the division.”
Lossless Intrinsic Image Decomposition via Learning Shading Feature Filtering
Sha, H., Han, Y., Xiao, Y., Liu, T., Liu, Y., & Song, W. (2025) · Computational Visual Media · doi:10.26599/cvm.2025.9450378 · Cited by 1
INVESTIGATION OF DIFFERENT METHODS OF THRESHOLD SELECTION FOR DCT-SSA FILTER
CHEREPANOV, I. (2025) · Herald of Khmelnytskyi National University Technical sciences · doi:10.31891/2307-5732-2025-349-67 · Cited by 1
Application of Optimized Adaptive Neuro-Fuzzy Inference for High Frame Rate Video Quality Assessment
Matulin, M., Mrvelj, Š., Periša, M., et al. (2025) · Applied Sciences · doi:10.3390/app15095018 · Cited by 1
Are Diffusion Models Effective Good Feature Extractors for MRI Discriminative Tasks?
Li, B., Sun, Z., Li, C., Kamagata, K., Andica, C., Uchida, W., Takabayashi, K., Guo, S., Zou, R., Aoki, S., et al. (2025) · Magnetic Resonance in Medical Sciences · doi:10.2463/mrms.mp.2024-0206 · Cited by 1
“5,7 Multi-Scale Structural Similarity Index Measure (MS-SSIM) 33 is used to assess diversity within distributions.”
DREFNet: Deep Residual Enhanced Feature GAN for VVC Compressed Video Quality Improvement
Das, T., & Choi, K. (2025) · Mathematics · doi:10.3390/math13101609 · Cited by 1
DeepFGS: Fine-Grained Scalable Coding for Learned Image Compression
Zhai, Y., Ma, Y., Tang, L., Jiang, W., & Wang, R. (2025) · DCC · doi:10.1109/dcc62719.2025.00034 · Cited by 12
“Our models are optimized with two quality metrics, i.e. MSE (mean squared error) and MS-SSIM (multiscale structural similarity) [37].”
Eye-tracking analysis of expertise-based visual behavior in 8K video quality assessment
Sugito, Y., & Tanaka, T. (2025) · ACM Symposium on Eye Tracking Research and Applications (ETRA) · doi:10.1145/3715669.3726818 · Cited by 1
Image Compression Model Based on Dynamic Convolution and Vision Mamba
Qiu, L., Bai, E., Wu, Y. P., Cao, Y., & Jiang, X. (2025) · IET Image Processing · doi:10.1049/ipr2.70080 · Cited by 1
Feng, Z., Liu, X., & Jung, C. (2025) · IEEE Access · doi:10.1109/access.2025.3573831 · Cited by 1
“We use a combination of weighted L 1 loss, L 2 loss [28], and multi-scale structural similarity index (MS-SSIM) loss [29] to effectively train the proposed NNLF.”
Tao, W., Luo, Y., Tong, J., et al. (2025) · Journal of Marine Science and Engineering · doi:10.3390/jmse13061085 · Cited by 1
Enhancing MIMO Radar Data Quality via Stereo-Camera-Guided Diffusion Models
Lee, H., Han, S., Kim, H., Lee, D., & Kim, D. (2025) · IEEE Sensors Journal · doi:10.1109/jsen.2025.3572911 · Cited by 1
A Swin Transformer-Based Approach to Semantic Transfer of Images
Gan, B., Wu, Z., Gou, J., et al. (2025) · doi:10.1109/eeice65049.2025.11033809 · Cited by 1
“MS-SSIM evaluates image quality by comparing images at multiple scales to obtain comprehensive structural information[12].”
Bannon, A., Rapeaux, A., Savolainen, O. W., Hadjipanayi, C., Yin, M., Chan, C. J., Lande, T. S., & Constandinou, T. G. (2025) · IEEE International Radar Conference · doi:10.1109/radar52380.2025.11031250 · Cited by 1
“…Signal to Noise Ratio (PSNR), Structural Similarity Index Measure (SSIM) [13], [14] and Multi-Scale Structural Similarity Index Method (MS-SSIM) [15], with results shown in Table I. Starting from the definition of Shannon entropy is is possible to estimate the efficiency of the proposed…”
BACF-Net: An Attention-Convolution Fusion Architecture for Learned Image Compression
Chang, C.-C., & Hsiao, H.-F. (2025) · IEEE International Symposium on Circuits and Systems proceedings · doi:10.1109/iscas56072.2025.11043965 · Cited by 1
Differentiable VMAF: A trainable metric for optimizing video compression codec
Li, J., Zhou, C., Chen, Y., et al. (2025) · IEEE International Symposium on Circuits and Systems proceedings · doi:10.1109/iscas56072.2025.11044280 · Cited by 1
“Additionally, we incorporate MS-SSIM [27] as the distortion term, setting α = { 256 , 512 , 1024 , 2048 } , and fine-tune for one epoch.”
Meng, C., An, P., Zhang, Q., & Yuan, F. (2025) · Optics Express · doi:10.1364/oe.565944 · Cited by 1
Du, Y., Matsumoto, K., Nagahara, S., Tabo, K., Hasegawa, M., & Hirobayashi, S. (2025) · IEEE Access · doi:10.1109/access.2025.3585189 · Cited by 1
“The structural similarity index measure (SSIM) [47], [48] was used to quantitatively assess spectral fidelity.”
InfluentialA New Deep Joint Source Channel Coding with GAN Discriminator for Wireless Image Transmission
Xu, Y., Shen, L., Qiu, S., et al. (2024) · International Conference on Ubiquitous Computing and Communications (IUCC) · doi:10.1109/iucc65928.2024.00070 · Cited by 1
“To prove this, we further use the MS-SSIM [21] index to measure the decoded images of these three methods.”
Lightweight three-stream encoder–decoder network for multi-modal salient object detection
Lu, J., Wang, T., Wan, B., Zhao, Q., Wang, S., Sun, Y., Zhou, Y., & Yan, C. C. (2025) · Journal of Visual Communication and Image Representation · doi:10.1016/j.jvcir.2025.104523 · Cited by 3
Wirabudi, A. A., Lim, S.-C., Lim, W., et al. (2025) · IEEE Access · doi:10.1109/access.2025.3586241 · Cited by 1
“Conventional evaluation metrics, such as the peak signal-to-noise ratio (PSNR) [39], multiscale structural similarity index (MS-SSIM) [40], and mean squared error (MSE) [46], are used to detect differences between reconstructed and original images.”
InfluentialHorváth, B. P., & Csáji, B. C. (2025) · Machine Learning · doi:10.1007/s10994-025-06814-0 · Cited by 1
“• Structural similarity index measure (SSIM) quantifies the similarity between two images by taking several parameters into account, such as contrast, luminance and structure (Wang et al., 2003).”
Rumman, M. I., Ono, N., Ohuchida, K., Nasution, A. K., Alqaaf, M., Altaf-Ul-Amin, M., & Kanaya, S. (2025) · Bioengineering · doi:10.3390/bioengineering12070764 · Cited by 1
“MS-SSIM is an extension of SSIM that evaluates images at multiple scales, enhancing its robustness to variations in viewing conditions [58].”
A Novel Approach for Image Inpainting using GAN
Mohod, A., & Nair, P. P. (2025) · 2025 6th International Conference on Control, Communication and Computing (ICCC) · doi:10.1109/iccc64910.2025.11077276
“• SSIM (Structural Similarity Index) [33]”
GHMSA-Net: Gated Hierarchical Multi-Scale Self-Attention for Perceptually-Guided AV1 Post-Processing
Zhao, B., Gwun, W., & Choi, K. (2025) · IEEE Access · doi:10.1109/access.2025.3596303 · Cited by 1
“…for early stopping or hyper-parameter tuning, thereby avoiding metric-induced bias [37] per sequence, capturing both local detail and global structure [38]. c. Video Multi-method Assessment Fusion (VMAF): A data-driven metric that regresses a bank of texture-and detail-aware features to MOS space.”
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.