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 3,701–3,800 of 6,047
Conditional GAN Approaches on Regression Labels: A State‐of‐the‐Art Review
Silva‐Silverio, A., Gómez-Gil, P., & Sánchez-Argüelles, D. (2025) · WIREs Data Mining and Knowledge Discovery · doi:10.1002/widm.70050 · Cited by 5
DA-Net: A Double Alignment Multimodal Learning Network for Point Cloud Quality Assessment
Wu, X., He, Z., Luo, T., et al. (2025) · IEEE Transactions on Image Processing · doi:10.1109/tip.2025.3640023 · Cited by 4
“In addition, to comprehensively compare the performance of projection-based FR methods, several standard FR-IQA metrics have also been introduced to evaluate the quality of projection map (using standard cube projection), which are including RMSE [58], PSNR [38], SSIM [39] and MS-SSIM [59].”
Huang, H., Ho, G. T. S., Geda, M. W., et al. (2025) · Scientific Reports · doi:10.1038/s41598-025-27959-x · Cited by 3
Cross-Architecture Knowledge Distillation for Deep Joint Source-Channel Coding
Dai, S., Chen, X., Deng, X., et al. (2025) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2025.3633266 · Cited by 3
Liu, X., Zhu, G., Li, Z., & Li, D. (2026) · Solar Energy · doi:10.1016/j.solener.2025.114311 · Cited by 4
Yan, J., Liu, Z., Chen, J., Xu, X., Chen, P., & Fang, Y. (2026) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2026.3651643 · Cited by 5
“Based on SSIM, numerous quality metrics have been proposed, such as MS-SSIM [16] (a multi-scale version) and IW-SSIM [17] (an information content weighting version).”
Bidirectional Motion-Enhanced Semantic Communication for Wireless Video Transmission
Zhang, Z., Yang, Q., He, S., & Shi, Z. (2026) · IEEE Internet of Things Journal · doi:10.1109/jiot.2026.3657245 · Cited by 3
“…of the bandwidth ρ is defined as We employ the peak signal-to-noise ratio (PSNR), the multi-scale structural similarity index measure (MS-SSIM) [45], and the learned perceptual image patch similarity (LPIPS) [46] to comprehensively evaluate the pixel, structural, and perceptual quality of the…”
A review of generative coverless image steganography based on diffusion models
Zhang, S., Xue, J., Fu, J., & An, D. (2026) · Neurocomputing · doi:10.1016/j.neucom.2026.133070 · Cited by 9
Dai, W., Cai, C., Sun, X., Long, L., & Lin, B. (2026) · Journal of Manufacturing Processes · doi:10.1016/j.jmapro.2026.02.025 · Cited by 4
Is JPEG AI Going to Change Image Forensics?
Cannas, E. D., Mandelli, S., Popović, N., Alkhateeb, A., Gnutti, A., Bestagini, P., & Tubaro, S. (2025) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw69036.2025.00167 · Cited by 15
Tweel, J. E. D., Ecclestone, B. R., Simmons, J. A. T., & Reza, P. H. (2026) · npj Imaging · doi:10.1038/s44303-026-00154-x · Cited by 5
“Relative performance was then quantified using two metrics: (1) Multi-Scale Structural Similarity Index (MS-SSIM) [65], which captures pixel-level structural similarity across scales, with higher values indicating better similarity; and (2) Deep Image Structure and Texture Similarity (DISTS) [66],…”
A deep representation learning model to predict response to vagus nerve stimulation
Suresh, H., Mithani, K., Li, V., Latypov, T. H., Warsi, N. M., Wong, S. M., Erdman, L., Kang, J., Germann, J., Gouveia, F. V., et al. (2026) · Nature Communications · doi:10.1038/s41467-026-71555-0 · Cited by 4
“We evaluated reconstruction fi delity using established metrics, including multiscale structural similarity 36 (MS-SSIM), Fré-chet Inception Distance 37 (FID), and mean squared error (MSE).”
A Generative Foundation Model for Chest Radiography
Ji, Y., Lin, D. Y., Wang, X., Zhang, L., Zhou, W., Ge, C., Chu, R., Yang, X., Zhao, J., Chen, J., et al. (2026) · NEJM AI · doi:10.1056/aioa2500799 · Cited by 10
FSAPF: A De-Scattering Framework With Stepwise Adjustment of Polarization Features
Lin, B., Fan, X., & Guo, Z. (2026) · IEEE Transactions on Image Processing · doi:10.1109/tip.2026.3690316 · Cited by 3
“The SSIM between the predicted image and the ground image (X, Y) is expressed as [37]: where µ x is the mean of image X, µ Y is the mean of image Y, σ x is the variance of image X, σ Y is the variance of image Y, σ XY is the covariance of X and Y, C 1 and C 2 are small constants to stabilize the…”
On Prediction of Image Denoising Expedience Using Neural Networks
Rubel, A., Rubel, O. S., & Лукин, В. В. (2018) · International Scientific-Practical Conference Problems of Infocommunications Science and Technology · doi:10.1109/infocommst.2018.8632050 · Cited by 3
Performance enhancement of a scanning electron microscope using a deep convolutional neural network
Panchal, S. V., Datar, S. S., & Gopinathan, U. (2022) · Measurement Science and Technology · doi:10.1088/1361-6501/ac4a19 · Cited by 3
Blurriness Guided Underwater Salient Object Detection
Peng, Y., Lin, Y., & Peng, W. (2021) · MTS/IEEE OCEANS Conference (OCEANS) · doi:10.23919/oceans44145.2021.9705721 · Cited by 3
Liu, S., Zhao, F., Jiang, H., et al. (2022) · AIP Advances · doi:10.1063/5.0114734 · Cited by 3
Defect Change Detection in Masonry Structures Using Image-Based 3D Photogrammetry
Chaiyasarn, K., Buatik, A., Mohamad, H., Zhou, M., & Poovarodom, N. (2023) · Journal of Architectural Engineering · doi:10.1061/jaeied.aeeng-1446 · Cited by 3
Extreme Learning Machine-Enabled Coding Unit Partitioning Algorithm for Versatile Video Coding
Jiang, X., Xiang, M., Jin, J., et al. (2023) · Information · doi:10.3390/info14090494 · Cited by 3
Caballero, H., Jiménez, V. M., & Ramos, M. A. (2023) · International Journal of Power Electronics and Drive Systems/International Journal of Electrical and Computer Engineering · doi:10.11591/ijece.v13i6.pp6335-6350 · Cited by 3
Image Transmission Based on Semantics Communication in Vehicle-Infrastructure Collaboration System
Li, J., Sheng, K., Ao, L., et al. (2023) · International Conference on Dependable Systems and Their Applications (DSA) · doi:10.1109/dsa59317.2023.00067 · Cited by 3
Magonov, J., Maier, J., Erath, J., et al. (2024) · Medical Physics · doi:10.1002/mp.16938 · Cited by 3
Mohammadi, S., Wu, Y., & Ascenso, J. (2023) · IEEE International Conference on Visual Communications and Image Processing (VCIP) · doi:10.1109/vcip59821.2023.10402691 · Cited by 5
“In many learning-based image compression solutions, metrics such as MSE or Multiscale structural similarity (MS-SSIM) [8] are used in the loss function and thus for optimization.”
MethodXianfeng, 王. W., Shiben, 刘. L., Jiandong, 田. T., Juanping, 赵. Z., yajing, 刘. L., & Chunhui, 郝. H. (2024) · Laser & Optoelectronics Progress · doi:10.3788/lop231770 · Cited by 3
Tyagi, S., Srivastava, S., & Sahana, B. C. (2024) · Multimedia Tools and Applications · doi:10.1007/s11042-024-19527-8 · Cited by 6
Suppressing High-Frequency Artifacts for Generative Model Watermarking by Anti-Aliasing
Li, Z., Liu, Y., Zhang, X., & Wu, H. (2024) · ACM Workshop on Information Hiding and Multimedia Security (IH&MMSec) · doi:10.1145/3658664.3659634 · Cited by 6
“Four common metrics were applied to evaluate the quality of the generated images, i.e., peak signal-to-noise ratio (PSNR), structural similarity (SSIM) [41], multi-scale structural similarity (MS-SSIM) [42], visual information fidelity (VIF) [30].”
MethodP‐119: Evaluating Seamlessness: A Quantitative Index for Transparent Tiled MicroLED Displays
Tsai, Y., Huang, M., Li, C.-H., et al. (2024) · SID Symposium Digest of Technical Papers · doi:10.1002/sdtp.17942 · Cited by 3
AdvLUT: Cloaking Geographic Location With Semantic-Based Adversarial 3-D Lookup Tables
He, Y., Ye, D., Tang, L., Liu, Z., & Chen, C. (2024) · IEEE Internet of Things Journal · doi:10.1109/jiot.2024.3470872 · Cited by 3
TLVC: Temporal Bit-rate Allocation for Learned Video Compression
Zhang, Y., Liu, M., Xu, C., Tang, Q., Yao, C., & Zhao, Y. (2024) · IEEE International Conference on Multimedia and Expo (ICME) · doi:10.1109/icme57554.2024.10688317 · Cited by 3
“Due to the proposed method being able to adapt to different bit-rate variations, we employ the commonly used Vimeo-90K septuplet dataset [2] to train only 2 models for MSE and MS-SSIM, respectively.”
BackgroundMethodEvaluating the Quality of Brain MRI Generators
Wu, J., Peng, W., Li, B., Zhang, Y., & Pohl, K. M. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-72117-5_28 · Cited by 4
“Popular metrics used for comparison are those commonly applied to 2D natural images, such as Multi-Scale Structural Similarity (MS-SSIM) [22], Fréchet Inception Distance (FID) [9], and Maximum-Mean Discrepancy (MMD) [7].”
MethodImage Enhancement Based on Histogram-Guided Multiple Transformation Function Estimation
Park, J., Vien, A. G., Pham, T. T., Kim, H., & Lee, C. (2024) · IEEE Transactions on Consumer Electronics · doi:10.1109/tce.2024.3476033 · Cited by 5
“For a quantitative evaluation, we use six quality metrics: PSNR, SSIM, MS-SSIM [57], FSIM [58], LPIPS [59], and CIE-DE2000 ((cid:8) E 00 ) [60].”
Enhanced Edge Detection for Image Segmentation and its Real-Time Implementation
Lourdu Jennifer, J. R., & P, J. V. R. S. (2024) · International Symposium on VLSI Design and Test (VDAT) · doi:10.1109/vdat63601.2024.10705714 · Cited by 3
“MS-SSIM adds texture and contrast data at multiple downscaled resolutions or scales, with brightness data at the maximum resolution level [20].”
BackgroundMorphFlow: Estimating Motion in In-Situ Tests of Concrete
Nogatz, T., Redenbach, C., & Schladitz, K. (2024) · Experimental Mechanics · doi:10.1007/s11340-024-01104-7 · Cited by 3
“Yet, the main reason for Wang et al. to propose SSIM [33] and later multi level SSIM [34] was that “classical” error measures like RMSE do not coincide well with human perception of similarity.”
GVVST: Image-Driven Style Extraction From Graph Visualizations for Visual Style Transfer
Song, S., Zhang, Y., Lin, Y., Qu, H., Wang, C., & Li, C. (2024) · IEEE Transactions on Visualization and Computer Graphics · doi:10.1109/tvcg.2024.3485701 · Cited by 7
“L SSIM represents Structural Similarity Index Measure (SSIM) loss [65], a loss function based on the Structural Similarity Index Measure, which assesses the similarity between two images by considering changes in structural information, luminance, and contrast.”
Multi-focus image registration based on optical flow tracking and Delaunay triangulation
Xia, X., Yang, D., Huo, S., Sun, J., & Xiang, H. (2024) · Signal Processing · doi:10.1016/j.sigpro.2024.109763 · Cited by 3
ImmerScope: Multi-view Video Aggregation at Edge towards Immersive Content Services
Chen, B., Guo, H., Wu, M., Yang, Z., Yan, Z., & Nahrstedt, K. (2024) · ACM Conference on Embedded Networked Sensor Systems (SenSys) · doi:10.1145/3666025.3699324 · Cited by 3
“Compared to systems based on the SOTA TSVCs and NSVCs on the WildTrack [12] dataset, ImmerScope achieves 64% and 78% average bandwidth savings at the same PSNR and the structural similarity index (SSIM) [83], respectively.”
A Semantic Conditional Diffusion Model for Enhanced Personal Privacy Preservation in Medical Images
Wang, S., Zhao, Z., Zhao, Y., Wang, L., Zhang, Y., Wang, J., Qiao, S., & Lv, Z. (2024) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2024.3511583 · Cited by 4
“…the quality and diversity of the generated medical images, we conduct a quantitative assessment using Fréchet Inception Distance (FID) [34], Mean Squared Error (MSE), Maximum Mean Discrepancy (MMD) [35], Multi-Scale Structural Similarity Index (MS-SSIM) [36], and Peak Signal-to-Noise Ratio (PSNR).”
Learned Image Compression for HE-Stained Histopathological Images via Stain Deconvolution
Fischer, M., Neher, P., Wald, T., Almeida, S. D., Xiao, S., Schüffler, P. J., Braren, R., GOETZ, M. A., Muckenhuber, A., Kleesiek, J., et al. (2025) · Lecture notes in computer science · doi:10.1007/978-3-031-77786-8_10 · Cited by 6
Challenges and Applications of Intrinsic Image Decomposition: A Short Review
Ulucan, D., Ulucan, O., & Ebner, M. (2025) · SN Computer Science · doi:10.1007/s42979-025-03659-1 · Cited by 3
“To merge the scores, we take inspiration from the study of Wang et al. where the MS-SSIM metric is designed based on human observations [67].”
Raymond, C., Yao, J., Clifford, B., et al. (2025) · American Journal of Neuroradiology · doi:10.3174/ajnr.a8566 · Cited by 3
Towards virtual painting recolouring using vision transformer on x-ray fluorescence datacubes
Bombini, A., Bofı́as, F. G.-A., Giambi, F., & Ruberto, C. (2025) · Machine Learning Science and Technology · doi:10.1088/2632-2153/adb937 · Cited by 3
“During training, the learning rate is changed dynamically via the P Y T ORCH native R EDUCE LRO N PL ATEU function by monitoring the selected metric, which is the Multi-Scale Structural Similarity Index (MS-SSIM) [87, 88] in the Lightning AI TORCHMETRICS implementation [89].”
MRI-Styled PET: A Dual Modality Fusion Approach to PET Partial Volume Correction
Lin, Y.-N. S., Huang, S., Tsai, C.-H., Wang, H.-W., Chung, M.-C., Gong, E., Hsiao, I., & Chen, K. T. (2025) · IEEE Transactions on Radiation and Plasma Medical Sciences · doi:10.1109/trpms.2025.3549617 · Cited by 3
“On the other hand, the ratio β of MR to PET was determined empirically as 2 The multiscale structual similarity index [MSSSIM, (10)] [26], an extension of the well-known SSIM [(9)] [27], was developed to evaluate image similarity across various resolutions.”
Astronomical Image Quality Assessment Based on Deep Learning for Resource-constrained Environments
Li, J., Zhang, X., Ge, J., Bai, C., Feng, G., Mu, H., Wang, L., Liu, C., Kang, Z., & Jiang, X. (2025) · Publications of the Astronomical Society of the Pacific · doi:10.1088/1538-3873/adb790 · Cited by 4
“Common full-reference assessment metrics include Root Mean Square ( RMS ) error, peak signal-to-noise ratio, multi-scale structural similarity index ( Wang et al. 2003 ) , and visual information fi delity ( Sheikh & Bovik 2006 ) .”
Elastographic magnetization prepared imaging with rapid encoding
Cerjanic, A. M., Diano, A., & Johnson, C. L. (2025) · Magnetic Resonance in Medicine · doi:10.1002/mrm.30482 · Cited by 3
“We chose average global stiffness as a comparison measure, as is common in brain MRE, and multi-scale structural similarity (MSSIM) 36 to compare distributions of properties.”
InfluentialFull-Reference Point Cloud Quality Assessment with Multimodal Large Language Models
Watanabe, R., Konno, T., Sankoh, H., Tanaka, B., & Kobayashi, T. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10889736 · Cited by 3
“Since SSIM [34] and MS-SSIM [35] are image quality assessment methods, their average evaluation results were measured across six views using the same images provided to the MLLMs.”
GSVC: Efficient Video Representation and Compression Through 2D Gaussian Splatting
Wang, L. E. I., Shi, Y., & Ooi, W. T. (2025) · doi:10.1145/3712678.3721876 · Cited by 3
Multiuser Content-Style Adaptive Semantic Communication for Image Transmission
Song, M., Ma, N., Liang, H., et al. (2025) · IEEE Internet of Things Journal · doi:10.1109/jiot.2025.3581393 · Cited by 3
“MS-SSIM is an image quality assessment method based on the human visual system [44].”
Gao, H., Wu, X., & Ding, X. (2025) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2025.3583211 · Cited by 3
“Considering RGT estimation as a regression problem with the goal to predict a RGT value for each pixel, we use the mean square error (MSE) and multiscale structural similarity (MS-SSIM) [36], [37] loss functions as labeled synthetic supervision to train the network.”
Lee, J., Loria-Salazar, S. M., Holmes, H. A., & Sayer, A. M. (2025) · Earth and Space Science · doi:10.1029/2025ea004338 · Cited by 4
“MS‐SSIM considers various metrics roughly corresponding to the luminance, contrast, and covariance of predicted and actual values within several differently sized sliding windows (Wang et al., 2003).”
CNN-based TEM image denoising from first principles
Chae, J., Hong, S., Kim, S., Yoon, S., & Kim, G. (2025) · Computational Materials Science · doi:10.1016/j.commatsci.2025.114146 · Cited by 10
Dong, B., Cao, B., Gui, G., et al. (2025) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2025.3601155 · Cited by 3
“The initial learning rate is 10 MS-SSIM is based on the similarity of images at different scales, which is more in line with human visual perception [30].”
Learned Lightweight Smartphone ISP with Unpaired Data
Arhire, A., & Timofte, R. (2025) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw67362.2025.00176 · Cited by 9
“The generated images achieve fidelity scores (PSNR, SSIM [39], MS-SSIM [38]) comparable to those obtained through paired training, while maintaining a favorable perceptual quality (LPIPS score [40]).”
ChatVis: Large Language Model Agent for Generating Scientific Visualizations
Peterka, T., Mallick, T., Yildiz, O., Lenz, D., Quammen, C., & Geveci, B. (2025) · IEEE Symposium on Large Data Analysis and Visualization · doi:10.1109/ldav68558.2025.00007 · Cited by 12
“The multi-scale structural similarity metric (MS-SSIM) [37] is among the most successful techniques for assessing perceptual similarity.”
Perceptual video quality estimation by regression with myopic experts
Giotsas, V., Deligiannis, N., Fisher, P., & Andreopoulos, Y. (2015) · International Conference on Quality of Multimedia Experience (QoMEX) · doi:10.1109/qomex.2015.7148115 · Cited by 3
“Some of the most successful ones are summarized below: • Multiscale-SSIM (MS-SSIM), an extension of the SSIM paradigm for still images [3], that has been shown to outperform the SSIM index and many other still image quality assessment algorithms [4].”
BackgroundVideo quality assessment method: MD-SSIM
Loh, W.-T., & Bong, D. B. L. (2015) · IEEE International Conference on Consumer Electronics - Taiwan (ICCE-Taiwan) · doi:10.1109/icce-tw.2015.7216904 · Cited by 3
Self-information weighting for image quality assessment
Peng, P., & Li, Z.-N. (2011) · International Congress on Image and Signal Processing (CISP) · doi:10.1109/cisp.2011.6100607 · Cited by 3
A Novel Approach to Minimize the Impact of Non Ideal Samples in Iris Recognition System
Arya, K. V., Gupta, A., Kumar, G., et al. (2012) · International Conference on Computer and Communication Technology · doi:10.1109/iccct.2012.77 · Cited by 3
“If the eyes are not properly opened, the images will have the occlusion which is caused due to eyelids and eyelashes.”
BackgroundA new image quality assessment database for tiled images
McFadden, S. B., & Ward, P. A. S. (2014) · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · doi:10.1117/12.2037969 · Cited by 3
An image quality metric based on the Harris corner
Yang, Y., Cheng, G., & Tu, D. (2012) · International Conference on Signal Processing · doi:10.1109/icosp.2012.6491719 · Cited by 3
A robust similarity measure for automatic inspection
Barkol, O., Kogan, H., Shaked, D., et al. (2010) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2010.5652547 · Cited by 3
Real-time Deblocked GPU Rendering of Compressed Volumes
Marton, F., Iglesias-Guitián, J. A., Iriberri, J. D., et al. (2014) · Eurographics · doi:10.2312/vmv.20141290 · Cited by 3
VHDL implementation of a No-Reference video quality metric using the Levenberg-Marquardt method
de Oliveira, M., da Silva, W. B., Fonseca, K. V. O., & Pohl, A. D. A. P. (2014) · IEEE International Symposium on Broadband Multimedia Systems and Broadcasting (BMSB) · doi:10.1109/bmsb.2014.6873474 · Cited by 3
Multimedia Content Security using Image Encryption
Choudhary, R., & Jb, A. (2014) · Cited by 3
Perceptual quality-aware active queue management for video transmission
Ghoreishi, S. E., Aghvami, A. H., & Martini, M. G. (2015) · IEEE International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC) · doi:10.1109/pimrc.2015.7343493 · Cited by 3
“In this paper, we use the multi-scale structural similarity (MS-SSIM) index [9] which provides a good approximation of user-perceived quality.”
MethodQuality Assessment of Mobile Videos
Cheon, M., Kim, S.-J., Chae, C., & Lee, J. (2014) · Springer eBooks · doi:10.1007/978-3-319-10368-6_4 · Cited by 3
Learning depth from a single image using visual-depth words
Kim, S., Choi, S., & Sohn, K. (2015) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2015.7351130 · Cited by 7
“Since the results using the proposed method only provides a relative depth order, we use the multi-scale structural similarity (MS-SSIM) [20] for an evaluation metric.”
MethodInfluentialAnalysis and Comparison of Objective Methods for Image Quality Assessment
Babkin, P., & Pavlov, Y. N. (2014) · Science and Education of the Bauman MSTU · doi:10.7463/0914.0726368 · Cited by 3
Ambiguity-based evaluation of objective quality metrics for image compression
Cheon, M., & Lee, J. (2016) · International Conference on Quality of Multimedia Experience (QoMEX) · doi:10.1109/qomex.2016.7498944 · Cited by 4
High fidelity imager emulator of measured systems
Haefner, D. P., & Teaney, B. P. (2017) · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · doi:10.1117/12.2260887 · Cited by 3
“Currently, there are very few objective metrics that can quantify the performance in the presences of a mostly non-linear and scene dependent algorithm [4,5].”
BackgroundA block level adaptive MV resolution for video coding
Ray, B., Jung, J., & Larabi, M. (2017) · IEEE International Conference on Multimedia and Expo (ICME) · doi:10.1109/icme.2017.8019451 · Cited by 3
“PSNR is used as the distortion metric for calculating BD Rate and MS-SSIM helps to report about perceptual visual quality [14].”
MethodStereoscopic Image Quality Assessment Based on Cyclopean Image
Lu, K., & Liu, X. (2016) · IEEE International Conference on Internet of Things (iThings) and IEEE Green Computing and Communications (GreenCom) and IEEE Cyber, Physical and Social Computing (CPSCom) and IEEE Smart Data (SmartData) · doi:10.1109/ithings-greencom-cpscom-smartdata.2016.158 · Cited by 3
“Such as Z. Wang.et proposed Structure Similarity (SSIM) [2] index and Mean Structural Similarity Index (MSSIM) [3] based on the idea of human visual system is sensitive with the structural information of image, A. Shnayderman.et proposed SVD-based Measure for local and global assessment (SVD) [4], and so on.”
BackgroundCompression performance of extended JPEG XT under various TMOs
Visavakitcharoen, A., Kobayashi, H., Shiota, S., et al. (2017) · IEEE Global Conference on Consumer Electronics (GCCE) · doi:10.1109/gcce.2017.8229372 · Cited by 3
A new quality assessment for Thangka image inpainting
Hu, W., Liu, Z., & Ye, Y. (2018) · Concurrency and Computation Practice and Experience · doi:10.1002/cpe.4671 · Cited by 3
Semantically Interpretable and Controllable Filter Sets
Prabhushankar, M., Kwon, G., Temel, D., & AIRegib, G. (2018) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2018.8451220 · Cited by 4
“Previous studies [25, 26, 27, 28, 29] in this field, including our own [11, 12] concentrated on and provided results on all the distortion levels together.”
BackgroundGPU Acceleration of the Most Apparent Distortion Image Quality Assessment Algorithm
Holloway, J., Kannan, V., Zhang, Y., Chandler, D. M., & Sohoni, S. (2018) · Journal of Imaging · doi:10.3390/jimaging4100111 · Cited by 4
“In [27], Okarma et al. presented a GPGPU technique for accelerating their combined video quality algorithm (a video quality assessment algorithm developed by Okarma using SSIM, MS-SSIM, and VIF [34]).”
MethodInfluentialStructure-preserving video super-resolution using three-dimensional convolutional neural networks
Liu, C., Li, X., Zhang, X., et al. (2019) · Journal of Electronic Imaging · doi:10.1117/1.jei.28.2.021007 · Cited by 3
Hepburn, A., Laparra, V., McConville, R., et al. (2020) · Proceedings of the Northern Lights Deep Learning Workshop · doi:10.7557/18.5124 · Cited by 3
“On the contrary, metrics like the one proposed by Laparra et al. are inspired by the early stages of the human visual cortex and show better performance in mimicking human perception than SSIM and MS-SSIM in different human rated databases.”
BackgroundFederated-Cloud Based Deep Neural Networks with Privacy Preserving Image Filtering Techniques
Choi, I., Song, Q., & Sun, K. (2019) · IEEE Conference on Dependable and Secure Computing (DSC) · doi:10.1109/dsc47296.2019.8937635 · Cited by 3
“• Multi-scale SSIM (MS-SSIM) [3]: SSIM conducted over multiple scales through a process of multiple stages of sub-sampling, reminiscent of multiscale processing in the early vision system.”
MethodDeep Learning Approach to Video Compression
Jacob, A., Pawar, V., Vishwakarma, V., et al. (2019) · IEEE Bombay Section Signature Conference (IBSSC) · doi:10.1109/ibssc47189.2019.8973035 · Cited by 3
HVS-Based Perceptual Color Compression of Image Data
Prangnell, L., & Sánchez, V. (2021) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp39728.2021.9414773 · Cited by 3
“In the experimental setup, an initial frame-level QP (iQP) is applied to all reference techniques in order to target scores of SSIM ≥ 0.99 and also MS-SSIM ≥ 0.99.”
MethodInfluentialAn Unsupervised Approach towards Varying Human Skin Tone Using Generative Adversarial Networks
Roy, D., Mukherjee, D., & Chanda, B. (2021) · Proceedings - International Conference on Pattern Recognition/Proceedings/International Conference on Pattern Recognition · doi:10.1109/icpr48806.2021.9412852 · Cited by 4
“To train the other half of this network which is responsible for the skin color estimation we used MS-SSIM [37] loss, which is a variant of SSIM [36].”
MethodJung, S., Kwon, H., Son, D.-M., et al. (2020) · IEICE Transactions on Information and Systems · doi:10.1587/transinf.2020edl0001 · Cited by 3
“The HDR-VDP-2 as an HDR visual detection predictor and the MS-SSIM based on multi-scale structural similarity are used for HDR image visibility and quality assessments [10], [11].”
MethodIntrospective Learning by Distilling Knowledge from Online Self-explanation
Gu, J., Wu, Z., & Tresp, V. (2021) · Lecture notes in computer science · doi:10.1007/978-3-030-69538-5_3 · Cited by 3
“The similarity can be measured with different metrics, such as cosine distance, multi-scale mean squared error [32], and Wasserstein distance [33].”
MethodAsymmetrically distorted 3D video quality assessment: From the motion variation to perceived quality
Fang, Y., Sui, X., Yan, J., Zuo, Y., Wang, J., & Li, Z. (2021) · Signal Processing · doi:10.1016/j.sigpro.2021.108031 · Cited by 3
Implementing Viewport Tile Extractor for Viewport-Adaptive 360-Degree Video Tiled Streaming
Jeong, J.-B., Lee, S., Kim, I., et al. (2021) · International Conference on Information Networking · doi:10.1109/icoin50884.2021.9333964 · Cited by 3
“To evaluate the performance of the proposed method, metrics, namely, peak signal-to-noise ratio (PSNR), video multimethod assessment fusion (VMAF) [8], multi-scale structural similarity (MS-SSIM) [9], and immersive video PSNR (IV-PSNR) [10], were used.”
MethodSelf-Supervised Simultaneous Multi-Step Prediction of Road Dynamics and Cost Map
Amirloo, E., Rohani, M., Banijamali, E., Luo, J., & Poupart, P. (2021) · 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/cvpr46437.2021.00839 · Cited by 5
“Finally, to push the predicted OGMs toward the target OGMs we use Structural Similarity Index Metrics (SSIM) [31].”
MethodInfluentialSparse Representations-based depth images quality assessment
Sebai, D., Sehli, M., & Ghorbel, F. (2021) · Visual Informatics · doi:10.1016/j.visinf.2021.02.004 · Cited by 3
Nizami, I. F., Akhtar, M., Waqar, A., Mann, A. B., & Majid, M. (2021) · Multimedia Tools and Applications · doi:10.1007/s11042-021-10672-y · Cited by 3
“…6 shows the performance comparison of the proposed methodology with sixteen state-of-the-art NR-IQA techniques, which include SSIM [42], MS-SSIM [43], FSIM [54], VIF [37], MAD [16], GMSD [46], DESIQUE [52], QAC [45], NIQE [29], ILNIQE [55], BPRI [26], BMPRI [27], GWH-GLBP [18], SISBLIM [10],…”
MethodInfluentialLearned Image Compression with Super-Resolution Residual Modules and DISTS Optimization
Suzuki, A., Akutsu, H., Naruko, T., Tsubota, K., & Aizawa, K. (2021) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw53098.2021.00215 · Cited by 3
“With our proposed method, we use loss function Ldm using DISTS [5] with Multi-Scale Structural SIMilarity (MSSSIM) [16] to estimate the score of human perceptual image quality in the training of the Gm.”
MethodTurbulent-degraded image restoration via improved principal component analysis method
Lü, Q., Zhuoma, S., Ge, B., et al. (2021) · Journal of Modern Optics · doi:10.1080/09500340.2021.1960445 · Cited by 3
“Table 2 shows PSNR, multiscale similarity SSIM (MS-SSIM) [ 25 ], feature similarity index (FSIM) [ 26 ] and visual information fidelity (VIF) [ 27 ] values for the outputs of four different methods.”
MethodYang, J., Wei, F., Bai, Y., et al. (2021) · Electronics · doi:10.3390/electronics10192434 · Cited by 3
Polynomial Image-Based Rendering for non-Lambertian Objects
Fachada, S., Bonatto, D., Teratani, M., & Lafruit, G. (2021) · IEEE International Conference on Visual Communications and Image Processing (VCIP) · doi:10.1109/vcip53242.2021.9675371 · Cited by 4
“For objective evaluation, we use three metrics (PSNR, IV-PSNR [33] and MS-SSIM [34]) averaged on all the rendered views.”
MethodRemoving Snow From a Single Image Using a Residual Frequency Module and Perceptual RaLSGAN
Sung, T. L., & Lee, H. J. (2021) · IEEE Access · doi:10.1109/access.2021.3126539 · Cited by 3
“𝐵𝐵 = 11 × 11 𝑀𝑀 For our convenience, we set .”
Colourlab Image Database: Geometric Distortions
Pedersen, M., & Amirshahi, S. A. (2021) · Color and Imaging Conference · doi:10.2352/issn.2169-2629.2021.29.258 · Cited by 3
Benchmarking Monocular Depth Estimation Models for VR Content Creation from a User Perspective
Dickson, A., Knott, A., & Zollmann, S. (2021) · International Conference on Image and Vision Computing New Zealand (IVCNZ) · doi:10.1109/ivcnz54163.2021.9653344 · Cited by 3
Tiny Generative Image Compression for Bandwidth-Constrained Sensor Applications
Körber, N., Siebert, A., Hauke, S., et al. (2021) · IEEE International Conference on Machine Learning and Applications (ICMLA) · doi:10.1109/icmla52953.2021.00094 · Cited by 3
Parameter space study of optimal scale-dependent weights in TV image denoising
Reyes, J. C. D. L., & Herrera, K. (2022) · Applicable Analysis · doi:10.1080/00036811.2022.2033231 · Cited by 3
“The Multi-Scale Structural Similarity Index (MSSIM) evaluates and combines the SSIM values across different resolutions (see [12]).”
MethodVideo Quality Assessment Based on Short-Term Memory
Fang, Y., Chen, W., Zhao, T., Xu, Y., & Chen, J. (2021) · KSII Transactions on Internet and Information Systems · doi:10.3837/tiis.2021.07.012 · Cited by 4
“They measure the perceptual quality of a given video by combining the quality of frame-level qualities, such as Peak Signal-to-Noise Ratio (PSNR) [6], Structural SIMilarity (SSIM) [7], Multi-Scale SSIM (MS-SSIM) [8], and so on.”
Digital design of regional characteristic apparel pattern driven by GAN
Chen, H., SHEN, L., ZHANG, X., et al. (2022) · Industria Textila · doi:10.35530/it.073.03.202117 · Cited by 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.