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 5,701–5,800 of 6,047
DSGV: Efficient Video Representation and Compression through Dynamic-Static 2D Gaussian Splatting
Luo, Z. (2025) · ACM International Conference on Multimedia in Asia (MMAsia) · doi:10.1145/3743093.3771035
“We evaluate video quality using PSNR and MS-SSIM [27].”
ST-VJSCC: Spatio-Temporal Video DeepJSCC Scheme for Adaptive Wireless Transmission
Zhou, X., Huang, D., Qi, Z., et al. (2025) · IEEE Transactions on Cognitive Communications and Networking · doi:10.1109/tccn.2025.3641521
“We evaluate the video reconstruction performance of the proposed ST-VJSCC model by PSNR and the perceptual metric MS-SSIM [41].”
A Hybrid Scheme for Face Video Compression
Tang, A., Zhang, Z. M., Zhu, C., Ling, J., Xie, R., & Song, L. (2025) · ACM Transactions on Multimedia Computing Communications and Applications · doi:10.1145/3783982
Live High-Fidelity Semantic Communication via Cross-Modal Fusion for Volumetric Video
Gong, T., Cao, Z., Liang, Z., Zhang, D., Wang, F., & Cui, S. R. (2025) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2025.3644826 · Cited by 1
“For communication robustness experiments, we additionally incorporate MS-SSIM [46] to measure structural similarity.”
A pre-log correction method based on dynamic approximation to reduce photon-starved deterioration
Liu, J., Chen, W., Jiang, H., et al. (2025) · Physical and Engineering Sciences in Medicine · doi:10.1007/s13246-025-01647-6
Wang, L., Lu, Y. M., Xu, S., Li, J., & Liu, J. (2025) · Journal of Lightwave Technology · doi:10.1109/jlt.2025.3646035
“WeevaluateFréchetInception Distance(FID)[43],KernelInceptionDistance(KID)[44], andMulti-ScaleStructuralSimilarityIndex(MS-SSIM)[45] atboth28 × 28 and 32 × 32 resolutions, as summarized in Table II.”
Hiding Local Manipulations on SAR Images: A Counter-Forensic Attack
Mandelli, S., Cannas, E. D., Bestagini, P., Tebaldini, S., & Tubaro, S. (2025) · IEEE Transactions on Image Processing · doi:10.1109/tip.2025.3643154 · Cited by 3
Yakymiv, V., Piskozub, Y., & Oliiarnyk, N. R. (2025) · Mathematical Modeling and Computing · doi:10.23939/mmc2025.03.914
Privacy-Preserving Robotic Perception for Object Detection in Curious Cloud Robotics
Antonazzi, M., Alberti, M., Bassot, A., Luperto, M., & Basilico, N. (2025) · IEEE Transactions on Robotics · doi:10.1109/tro.2025.3613551 · Cited by 2
“The potential privacy breach by the attacker is assessed utilizing the multiscale structural similarity index measure (MS-SSIM) as defined in [51], with reconstructed images ˆ x serving as inputs.”
Imai, T., & Mizutani, T. (2025) · International Workshop on Advanced Ground Penetrating Radar · doi:10.1109/iwagpr65621.2025.11108984
“The loss function for inversion L main was the sum of the two multiscale structural similarity (MS-SSIM)[12] and mean absolute error (MAE).”
2D Gaussian Splatting With Pre-Trained Dictionaries for Image Compression
Gao, Y., Zhang, X., & Zhang, G. (2025) · IEEE Signal Processing Letters · doi:10.1109/lsp.2025.3619027
“We use PSNR and MS-SSIM [23] to measure the image quality, and qualify the bitrate by bit per pixel (bpp).”
Cross-Scale Style-Guided Enhancement for Underwater Remote Sensing Imagery
Zhou, J., Zhang, D., Han, X., et al. (2025) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2025.3629321
Towards complex seismic layers: A vision Transformer approach to full waveform inversion
Li, Y., Fan, A., Pan, S., Song, G., & Min, F. (2025) · Computational Geosciences · doi:10.1007/s10596-025-10398-y · Cited by 1
“In particular, MAE and MSE are employed to measure the pixel-wise error, whereas SSIM, MS-SSIM and LPIPS are used to indicate the similarity of images.”
Using Diffusion Transformers to Generate Synthetic Diffusion Scalar Maps for Data Augmentation
Chattopadhyay, T., Jagad, C., Senthilkumar, P., Thomopoulos, S. I., Villalón-Reina, J. E., & Thompson, P. M. (2025) · International Symposium on Medical Information Processing and Analysis (SIPAIM) · doi:10.1109/sipaim67325.2025.11283192
“We evaluate synthetic image quality using Multi-Scale Structural Similarity (MS-SSIM) [12] and Maximum Mean Discrepancy (MMD) [13], ensuring pixel-level accuracy and distributional alignment with real data.”
REMAC: Reference-Based Martian Asymmetrical Image Compression
Ding, Q., Xu, M., Li, S.-X., et al. (2026) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2025.3649222
“As shown, the proposed REMAC achieves superior compression performance across PSNR, MS-SSIM [43], and LPIPS [44] metrics, while maintaining the third-lowest encoder complexity.”
Visual Integrity in the Age of AI: An Evaluation of DLSS and DLAA in Geospatial Visualization
Beregovyi, K., & Butkiewicz, T. (2025) · IEEE Visualization and Visual Analytics (VIS) · doi:10.1109/vis60296.2025.00064
“Based on a survey of image quality methods for compressed images [6], Butteraugli [1] and Multi-Scale Structural Similarity Index (MS-SSIM) [13] were chosen as complimentary standardized comparison metrics.”
Li, H., Zhang, B., Liu, Y., et al. (2025) · Electromagnetics · doi:10.1080/02726343.2025.2604017
“Next, we add the structural similarity (SSIM) to evaluate the performance of all methods, whose definition can be found in (Wang et al. 2003).”
Minigs: efficient 3D Gaussian splatting with full factors weighted pruning for scene representation
Qing, Y., Dongye, X., Guo, H., et al. (2026) · Multimedia Systems · doi:10.1007/s00530-025-02104-w
“This selection covers a diverse range of environments, including both indoor and outdoor settings, as well as natural and artificial scene types.”
Liu, X., Lei, H., Niu, G., Ye, S., Zheng, L., Liu, J., Zhou, Y., Liu, F., Lv, Y., & Gu, Y. (2026) · Lecture notes in computer science · doi:10.1007/978-981-95-5740-0_20
Energy-Efficient Approximate Prewitt Filter Design
Masadeh, M., Idrais, A. R., & AlShorman, O. (2025) · doi:10.1109/icm66518.2025.11322445
“The multiscale structural similarity (MS-SSIM) index [21], which compares the input image to the reference image and displays the results in Table IV, is used to determine the quality of the output images when applying different designs.”
Full-reference image quality assessment metrics with signal strength quantification capabilities
Maiseli, B. J. (2026) · The Imaging Science Journal · doi:10.1080/13682199.2026.2612885 · Cited by 1
“…an image from the TID2008 dataset labelled I17 (reference image) along with its versions of distortions (Figures 7) and ranked with respect to their consistency with MOS (subjective score) (Table 8): FSIM & FSIM C [46], MS-SSIM [47], VIF [48], SSIM [37], VSNR [49], IFC [50], NQM [2,51 values.”
BELE: Blur Equivalent Linearized Estimator
Giannitrapani, P., Claudio, E. D. D., & Jacovitti, G. (2026) · IEEE Transactions on Image Processing · doi:10.1109/tip.2026.3651959 · Cited by 1
Wang, X., Liu, X., Xie, X. N., Zhang, K., Cai, Q., Tong, X., Gong, Z., & Li, K. (2025) · IEEE International Conference on Parallel and Distributed Systems (ICPADS) · doi:10.1109/icpads67057.2025.11323172
“We employ the peak signal-to-noise ratio (PSNR) and multi scale structural similarity index measure (MS-SSIM) [18] to evaluate the quality of reconstructed images from the legitimate receiver or the attacker.”
Liu, X., Chen, W., & Shan, L. (2025) · International Conference on Big Data, Artificial Intelligence and Internet of Things Engineering (ICBAIE) · doi:10.1109/icbaie66852.2025.11326614
ROI-Based Image Transmission With Channel Feedback for Semantic Communications
Teng, W., Han, W., Wu, W., et al. (2026) · IEEE Wireless Communications Letters · doi:10.1109/lwc.2026.3654189
“As for MS-SSIM, the corresponding loss function can be expressed as where MS-SSIM ( · , · ) is defined in [16].”
InfluentialDetection of low-density foreign objects in confectionery products using sub-terahertz technology
Papadopoulos, K., Brüninghaus, J., & Jelali, M. (2026) · Journal of Food Measurement & Characterization · doi:10.1007/s11694-025-03937-9
A Lightweight Codebook-Assisted Semantic Communication Architecture for UAV Image Transmission
Liu, C., Zhou, L., Deng, X., et al. (2025) · doi:10.1109/cloudcom67567.2025.11331466
“We adopt the peak signal-to-noise ratio (PSNR) and the multi-scale structural similarity index (MS-SSIM) [20] as performance metrics.”
Fallatah, S., Lisitsa, A., & Grasso, F. (2025) · International Conference on Federated Learning Technologies and Applications (FLTA) · doi:10.1109/flta67013.2025.11336523
“• Structural Similarity Index (SSIM): Measures perceptual similarity by comparing luminance, contrast, and structural information between the original and reconstructed images [38].”
An Attention-Guided Interactive Network for Infrared and Visible Image Fusion
Zhang, M., Cao, Y., & Zhang, Z. (2025) · IEEE Advanced Information Management, Communicates, Electronic and Automation Control Conference (IMCEC) · doi:10.1109/imcec66174.2025.11331866
“To further promote the synergy between fusion results and high-level vision tasks, Tang et al. [23] proposed a semantic-aware loss function, making the fused images more conducive to subsequent visual analysis.”
StyleToonEdit: Face Cartoon Stylization and Controllable Generation Based on StyleGAN Inversion
Song, Y., Jia, X., & Liang, Y. (2026) · Lecture notes in computer science · doi:10.1007/978-981-95-5628-1_3
Mask Optimization with Auxiliary Lines for Image Inpainting
Du, W., Urano, Y., & Sari, I. N. (2025) · International Conference on Systems and Informatics (ICSAI) · doi:10.1109/icsai68704.2025.11345855
“The following evaluation-based methods are used as quantitative measures to verify the effectiveness of our proposal: Peak Signal to-Noise Ratio(PSNR) [14], Structural Similarity(SSIM) [15] and Multi-Scale Structural Similarity(MSSSIM) [16].”
Mixed-Reference Quality Assessment for Novel View Synthesis Scenes
Hao, J., Yang, S., Duan, H., Zhang, D., Han, Y., Yuan, L., & Zhai, G. (2026) · Lecture notes in computer science · doi:10.1007/978-981-95-5682-3_8
Kyu, P. M., & Woraratpanya, K. (2026) · IEEE Access · doi:10.1109/access.2026.3657196 · Cited by 1
“Metrics such as SSIM [12], MS-SSIM [13], and IW-SSIM [14] emphasize structural similarity but often fail to detect perceptual artifacts like unnatural textures or color distortions.”
He, L., Wu, Y., Wu, J., Wang, D.-H., Zhu, S., Zhang, X., & Liu, M. (2026) · Lecture notes in computer science · doi:10.1007/978-981-95-5631-1_2
Single Image Reflection Separation by Using Reflection and Refraction Estimations
Liu, T.-J., Chan, U.-I., & Liu, K.-H. (2025) · IEEE International Conference on Systems, Man and Cybernetics · doi:10.1109/smc58881.2025.11343404
“SSIM (Structural Similarity) Loss [30]: This loss measures the similarity between the predicted image T pre and the ground truth image T .”
Phase-Guided Cross-Frequency Integration Network for ISAR and Optical Image Fusion
Wang, Z., Liu, L., Zhang, Z., Du, R., Zhou, W., Zhou, M., Cai, J., & Zhou, F. (2026) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2026.3657411
“…(SCD) [61] reflects the degree of information preservation and transmission from the source inputs to the fused output; 4) SSIM [62] and 5) MS-SSIM [63] quantify the structural similarity between the source and fused images; 6) AG [64] and 7) EI [65] indicate the richness of texture information in…”
Self-citationHung, K.-J., & Perng, J.-W. (2026) · The International Journal of Advanced Manufacturing Technology · doi:10.1007/s00170-026-17492-0 · Cited by 1
AI-based Denoising and Interpolation of Magnetic UXO Data
Kovalenko, M., Przewozny, D., Chojecki, P., et al. (2025) · Conference proceedings/Conference proceedings - IEEE International Conference on Systems, Man, and Cybernetics · doi:10.1109/smc58881.2025.11342490
“Inspired by Zhao et al. [36], who combine Multiscale Structural Similarity Index (MS-SSIM) [37] and Mean Absolute Error (MAE), we use a similar dual-loss strategy but replace MAE with Mean Squared Error (MSE) to better match the floating-point characteristics and sharp dipole peaks of magnetometer…”
Hierarchical Conditional Guidance Diffusion Model for Perceptual Image Compression
Ji, Z., Qin, J., Qin, P., Chai, R., Zeng, J., & Lin, Z. (2025) · Conference proceedings/Conference proceedings - IEEE International Conference on Systems, Man, and Cybernetics · doi:10.1109/smc58881.2025.11342965
“Meanwhile, we also employed Peak Signal-to-Noise Ratio ( PSNR ) and Multi-Scale Structural Similarity Index Measure ( MS-SSIM ) [27] as quantitative indicators to evaluate content fidelity.”
Frequency-Decomposed Interaction Network for Stereo Image Restoration
Tian, X., Xie, J., Xu, R., Nie, J., Cao, J., Pang, Y., & Li, X. (2026) · IEEE Transactions on Image Processing · doi:10.1109/tip.2026.3658219 · Cited by 2
“In contrast, for the stereo image raindrop removal task, following [8], we utilize PSNR, MS-SSIM [67], and LPIPS [68] to evaluate restoration quality.”
On the Suitability of Perceptual Quality Metrics for Learning-Based Screen Content Compression
Dogaroglu, H. B., You, H., Boev, A., et al. (2025) · International Symposium on Multimedia · doi:10.1109/ism66958.2025.00053
“The final set of metrics includes PSNR, NLPD [16], SSIM [9], MS-SSIM [13], FSIM [14], and VIF [15].”
Dual Attention XYDeblur: Attention-Guided Divide-and-Conquer Strategy for Image Deblurring
Shin, J. S., & Lee, H. S. (2026) · IEEE Access · doi:10.1109/access.2026.3660952 · Cited by 1
“…XYDeblur surpasses recent methods, such as MRD-Net [37], demonstrating a 0.94 dB improvement in peak signal-to-noise ratio (PSNR) [71], a 0.0087 increase in structural similarity (SSIM) [71], and a 0.0071 enhancement in multi-scale structural similarity (MS-SSIM) [72] on the GoPro dataset [17].”
InfluentialMetrics for the Perceptual Evaluation of Lighting Estimation AlgorithmsPreprint
Giroux, J., Hold-Geoffroy, Y., Vázquez-Corral, J., et al. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6177599
EfficientSINet-B4: Enhancing Crowd Counting with EfficientNet and Shunting Inhibition Mechanism
Reddy, M. S., Reddy, Y., Mani, L. C., et al. (2025) · IEEE Global Conference for Advancement in Technology (GCAT) · doi:10.1109/gcat66372.2025.11368473
“Others use segmentation [16], edge guides [17], or look at the image in a smart way [18].”
Visually enriching and comparing runtime performance of visualization pipelines
Tarner, H., Gralka, P., Reina, G., et al. (2026) · Journal of Visualization · doi:10.1007/s12650-025-01091-4
Liu, Y., Xu, M., Shi, Y., et al. (2026) · Lecture notes in computer science · doi:10.1007/978-981-95-6960-1_34
PLGMNet: Parallel Local-Global Mamba Network for Real-Time Steel Surface Defect Detection
Zhang, J., Li, C., Zhou, X., et al. (2026) · Lecture notes in computer science · doi:10.1007/978-981-95-5758-5_26
Enhancing Learned Image Compression through Side Information AutoregressionPreprint
He, H., Shi, Y., Ren, J., et al. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6184936
Creating a Neural Network Noise Reduction Metric Based on Human Color Perception
Abramov, D. E., Korchagin, S. A., Zaichenkova, E. E., et al. (2025) · Journal of Communications Technology and Electronics · doi:10.1134/s1064226925700470
FCFNet: A Frequency-Domain-Guided Cross-Modal Feature Fusion Network for Cloud Removal
Wu, C., Xu, F., Li, X., et al. (2026) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2026.3662395
“The L MS-SSIM loss preserves multiscale structural coherence by evaluating luminance, contrast, and pattern correlations across spatial resolutions [61] Furthermore, given the critical role of frequency-domain modeling in our framework, the L fre loss explicitly constrains spectral distribution…”
Spatiotemporal Dual Priors Guided Diffusion for Remote Sensing Spatiotemporal Fusion
Zhu, B., Song, H., Zhang, K., & Liu, G. (2026) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2026.3664495 · Cited by 1
“ˆ: c = TCEM (Encoder 1 ( C 1 ) − Encoder 1 ( C 2 )) R In the design of the loss function, to make PENet focus more on the coarse-to-fine mapping, i.e., the di ff erence between the fine and coarse images, we add a structural loss [57] on top of the pixelwise loss [58].”
Adaptive structure-guided impedance inversion using implicit sparse constraint
Li, H., Cui, Y.-A., Wang, P., et al. (2026) · Petroleum Science · doi:10.1016/j.petsci.2026.02.014
TFFN: Three-Branch Feature Fusion Network for Stereoscopic Omnidirectional Image Quality Assessment
Liu, Y., Li, S., Fan, D., Duan, H., Jing, P., Yue, G., & Zhai, G. (2026) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2026.3664949
“The FR metrics for comparing include six 2D IQA-based metrics (PSNR, IW-PSNR [77], SSIM [25], MS-SSIM [27], FSIM [78] and VSI [79]), three 2D OIQA-based metrics (S-PSNR [80], WS-PSNR [81] and CPP-PSNR [82]), four 3D IQA metrics (Chen et al. [53], W-SSIM [44], W-FSIM [44] and SINQ [83]), and five 3D…”
HoloQA : Full Reference Video Quality Assessor of Rendered Human Avatars in Virtual Reality
Saha, A., Chen, Y.-C., Häne, C., Bazin, J., Katsavounidis, I. X., Chapiro, A., & Bovik, A. C. (2026) · IEEE Transactions on Image Processing · doi:10.1109/tip.2026.3663930
“A variety of perceptually motivated statistical feature extractors are available, including structural similarity [5], [6], [56], entropy di ff erencing [16], [17], [18], and contrast masking [19], [22].”
InfluentialSelf-citationLow-light enhancement on nighttime unmanned aerial vehicle (UAV) imagery for surveying vacant houses
Yu, B., Iizuka, K., Akiyama, Y., Oguchi, T., & Sugita, S. (2026) · Remote Sensing Applications Society and Environment · doi:10.1016/j.rsase.2026.101931
Self-Supervised Representation Learning with Diffusion-Based Refinement for Image Rectification
Kumari, P., & Das, S. (2025) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw69036.2025.00369
Liu, S., He, Z., Jiang, G., Jin, C., Zhou, W., Zhu, L., & Luo, T. (2026) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2026.3667351 · Cited by 2
“Torlig et al. [10] tested di ff erent IQA methods on such projection maps, such as SSIM [32], MS-SSIM [33], and VIFP [34].”
HFRW: High Fidelity and Robust Watermarking Using Deep Reinforcement Learning
Ping, P., Jiang, R., Guo, B., Yang, X., & Xu, F. (2026) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2026.3667156
“Here, we use a dual-component loss function: MSE and (MS-SSIM) [44].”
Kohara, Y., Nakayama, R., Nagata, M., et al. (2026) · IEEE Access · doi:10.1109/access.2026.3666877
“The registration performance of the proposed and comparative methods was quantitatively assessed using several widely adopted metrics: root mean square error (RMSE), multi-scale structural similarity index measure (MS-SSIM) [24], and Dice similarity coefficient (Dice) [25].”
Intelligent Image Processing System for UAV Navigation
Holovachov, A. (2025) · 2025 IEEE 8th International Conference on Methods and Systems of Navigation and Motion Control (MSNMC) · doi:10.1109/msnmc61130.2025.11399715
Towards Edge Deployment: An Ultra Lightweight Encoder for Learned Image Compression
Diao, P., Liu, L., Shen, Q., et al. (2025) · IEEE International Conference on Visual Communications and Image Processing (VCIP) · doi:10.1109/vcip67698.2025.11396826
“The rate is defined as the bits per pixel (bpp) of the compressed image and the distortion is calculated by PSNR and MS-SSIM [29] between the original image and the reconstructed image.”
Situational Quality: How Context Shapes Image Quality Perception and Affects Objective Measures
Zizien, A., Sarr, S. M. D., & Fliegel, K. (2025) · IEEE International Conference on Visual Communications and Image Processing (VCIP) · doi:10.1109/vcip67698.2025.11396842
“The predictive performance of these context-specific MPSNR models was then benchmarked against other widely used and well performing conventional IQA metrics, namely PSNR, SSIM [23], MS-SSIM [24], FSIM [25], VIF [26], and learning-based PieAPP [27] and TOPIQ FR [28].”
One-Step Diffusion for Perceptual Image Compression
Jia, Y., Wei, H., Zhou, Y., & Ge, C. (2025) · IEEE International Conference on Visual Communications and Image Processing (VCIP) · doi:10.1109/vcip67698.2025.11396876 · Cited by 2
A Deep Unrolling DBIM for Microwave Inverse Imaging With Inhomogeneous Background
Zhou, T., Lan, P., Zhou, X., Huang, J., Huo, J., & Peng, T. (2026) · IEEE Transactions on Antennas and Propagation · doi:10.1109/tap.2026.3666200
“The SSIM is a well-known quality assessment that is used to measure the similarity between two images [49].”
Li, H., Liu, S., Wang, D., Zhao, Y., Wei, X., Xian, W., Yan, J., He, X., Yu, F. Y., Luo, J., et al. (2026) · IEEE Transactions on Broadcasting · doi:10.1109/tbc.2026.3666657
“This includes prominent traditional methods such as PSNR, SSIM [56], MS-SSIM [57], VIF [58], MAD [52], FSIM [59], VSI [60], GMSD [61], and NLPD [62].”
Improving the Transferability of Adversarial Examples Through Spatial-Based Color Modification
Zhou, Z., Wang, Z., Zhang, X., & Feng, G. (2026) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2026.3668693
“Then optimize the back-bone using both the L 1 loss function and the MS-SSIM [68] loss function: where ω l 1 and ω ms − ssim representthehyperparametersthatcon-trol the two components of the loss function.”
Self-citationRestoration of Non-Uniform Motion-Blurred Star Images Based on Dynamic Strip Attention
Han, J., Niu, Z., & He, J. (2026) · Journal of Imaging · doi:10.3390/jimaging12030103 · Cited by 1
“Structural Similarity Index Measure (SSIM) calculates the similarity by comparing three key features between images: luminance, contrast, and structure [40], which is more consistent with the intuitive perception of the human eye.”
Predicting Satisfied User and Machine Ratio for Compressed Images: A Unified Approach
Zhang, Q., Wang, S., Zhang, X., Meng, X., Ma, S., & Gao, W. (2026) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2026.3668717 · Cited by 1
“We carefully select 14 representative FR-IQA models for labeling, which are SSIM [17], MS-SSIM [18], VIF [19], CW-SSIM [20], FSIM [21], LPIPS-Alex and LPIPS-VGG [2], PieAPP [22], DISTS [3], CKDN [23], ST-LPIPS-Alex and ST-LPIPS-VGG [24], and TOPIQ and TOPIQ-Pipal [4].”
Efficient Coding Parameters Optimization for Rate Control in Versatile Video Coding
Zhao, Z., He, X., Bi, X., et al. (2026) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2026.3668683
“Furthermore, to assess the suitability of the proposed algorithm for human visual perception, BD-rate calculations employing the Multiscale Structural Similarity Index (MS-SSIM) [45] are also utilized.”
Rate Control for 360$^{\circ }$ Versatile Video Coding Based on Visual Gaze Mechanism
Zhao, Z., Wang, M., Sui, X., Chen, P., He, X., & Wang, S. (2026) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2026.3668626
“In addition, in order to ascertain the suitability of the proposed model for human eye perception, BD-rate based on Structural Similarity Index (SSIM), Muti-Scale Structural Similarity Index (MS-SSIM) [40], and Video Multimethod Assessment Fusion (VMAF) [41] are also employed as evaluation measures.”
DAA-Font: Cross-Lingual Font Generation With Dual-Branch Encoding and Multi-Level Axial Attention
Yan, X., Mu, D., Zhang, Y., et al. (2026) · IEEE Access · doi:10.1109/access.2026.3670104
“The Structural Similarity Index (SSIM) [51] and Multi-Scale SSIM (MS-SSIM) [52] were used to measure luminance, contrast, and texture consistency across multiple scales.”
Liao, X., Chai, X., Chen, H., Jing, W., Shao, F., & Jiang, Q. (2026) · IEEE Transactions on Broadcasting · doi:10.1109/tbc.2026.3668512 · Cited by 1
“For 2D FR-IQA, these include: PSNR [52], which measures pixel-wise error; SSIM [40] and MS-SSIM [53], which evaluate structural similarity; FSIM [54], based on phase congruency and gradient magnitude; and VIF [55], based on information fidelity.”
Senthilkumar, T., Vinodkumar, A., & Sriramakrishnan, P. (2026) · Mathematical Methods in the Applied Sciences · doi:10.1002/mma.70626
A No-Reference Framework for Animated Video Quality Prediction in Consumer Multimedia Applications
Mogasati, R. S., Ganavdiya, S., Poreddy, A. K. R., Appina, B., & Pachori, R. B. (2026) · IEEE transactions on consumer electronics · doi:10.1109/tce.2026.3671988
A foundation model for multi-task cross-distribution restoration of fluorescence microscopy images
Lu, Q., Liu, X., Feng, Q., Zeng, S., & Cheng, S. (2026) · Nature Communications · doi:10.1038/s41467-026-70307-4 · Cited by 3
“Across the majority of internal data-sets, FluoResFM attained higher average peak signal-to-noise ratio (PSNR), multi-scale structural similarity index measure (MS-SSIM), zero-normalized cross-correlation (ZNCC), resolution-scaled Pearson coef fi cient (RSP) values and lower resolution-scaled error (RSE) values (Fig.”
InfluentialBit Allocation Transfer for Perceptual Quality Enhancement of VVC Intra Coding
Yang, R., & Bajić, I. V. (2025) · Picture Coding Symposium · doi:10.1109/pcs65673.2025.11417510
“We train three perceptual variants: 1-SSIM, 1-MS-SSIM, and LPIPS.”
InfluentialMonte Carlo Synthetic Data Generation for Radiograph Denoising
Roginek, E. W., Kulesza, J., & Zhao, Y. (2025) · IEEE ... International Conference on Data Mining workshops · doi:10.1109/icdmw69685.2025.00016
Scene-Adaptive Neural Video Compression with Dynamic Resource Allocation
Yang, Y.-W., & Hsiao, H.-F. (2025) · Picture Coding Symposium · doi:10.1109/pcs65673.2025.11417566
“We evaluate rate-distortion performance via BPP-PSNR and BPP-MS-SSIM [14] curves, plus encoding time to validate efficiency gains.”
A Hybrid Framework Integrating End-to-End Learned Image Codec with Conventional Codec
Zou, N., Hallapuro, A., Cricri, F., et al. (2025) · Picture Coding Symposium · doi:10.1109/pcs65673.2025.11417647
“The selected perceptual quality metrics include: Multi-Scale Structural Similarity Index (MS-SSIM) [22], Visual Information Fidelity (VIF) [23], Feature Similarity Index (FSIM) [24], Normalized Laplacian Pyramid Distance (NLPD) [25], Information-Weighted Structural Similarity Index (IW-SSIM) [26],…”
Fast localization and rectification of documents folded into thirds
Ershov, A., Tropin, D., & Nikolaev, D. P. (2025) · Computer Optics · doi:10.18287/coj1755
“…the original grayscaled image that was printed out with the following standard metrics: multiscale structural similarity index (SS = 1 – MS SSIM, [55]), local distortion (LD, [22]), absolute Levenstein distance between a string extracted from the input image and reference image, also called edit…”
TreeNet: A Light Weight Model for Low Bitrate Image Compression
Panda, M. P., Makkena, P. R., Prativadibhayankaram, S., et al. (2025) · Picture Coding Symposium · doi:10.1109/pcs65673.2025.11417522
“For training the models following loss function was utilized, ˆ x = g s ( g eb ( g a ( x ))) where, R represents rate, λ 1 and λ 2 are Lagrangian multipliers, and Ψ indicates factorized prior.”
Huang, M., Zhang, Y., Chandler, D. M., Leszczuk, M., & Farias, M. C. Q. (2026) · Journal on Image and Video Processing · doi:10.1186/s13640-026-00687-6 · Cited by 1
Enhancing Content Representation for AR Image Quality Assessment Using Knowledge Distillation
Sekhri, A., Amirshahi, S. A., & Larabi, M. (2026) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2026.3673101 · Cited by 1
“The metrics tested include PSNR, NQM [49], SSIM [33], MS-SSIM [34], VIF [50], IW-MSE, IW-PSNR, IW-SSIM [35], FSIM [36], GSI [51], GMSD [52], GMSM [52], PAMSE [53], LTG [54], and VSI [55].”
Kwong, N.-W., Chan, Y., Huang, Z., & Tsang, S. (2026) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2026.3674552
“ForVQA methods: In the early development of VQA, IQA methods [9], [10], [11] were adapted to extract visual quality features and evaluate the quality of frames, and were then employed using machine learning or linear functions for temporal pooling.”
EditHF-1M: A Million-Scale Rich Human Preference Feedback for Image Editing
Xu, Z., Duan, H., Ji, Z., Zhang, X., Liu, Y., Min, X., Gu, K., Zhang, J. D., Xu, S., Chen, J., et al. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-37098-3_16 · Cited by 5
Power-law Indices of EUV Intensity Power Spectrum in Flaring Coronal Active Regions
Zhong, S., Kolotkov, D. Y., & Nakariakov, V. M. (2026) · The Astrophysical Journal · doi:10.3847/1538-4357/ae48e5
A multi-conditional prior-guided dual spiral CNN-Mamba network for face sketch-photo synthesis
Liu, Y., Xu, M., Long, Y., et al. (2026) · Journal of King Saud University: Computer and Information Sciences · doi:10.1007/s44443-026-00652-3
“To comprehensively evaluate the performance of our proposed model, we adopt seven widely used image quality metrics: The Learned Perceptual Image Patch Similarity (LPIPS) (Zhang et al. 2018) metric, Fréchet Inception Distance (FID) (Heusel et al. 2018), Multi-Scale SSIM (MSSIM) (Wang et al. 2003), visual information fi…”
Kim, W., Lee, J., Shin, W., & Lee, J. (2025) · Globecom · doi:10.1109/globecom59602.2025.11432246
“We use peak-signal-to-noise ratio (PSNR) [15] and multiscale structural similarity index measure (MS-SSIM) [16] as the performance metric for numerical comparison.”
ProGVC: Progressive-Based Generative Video Compression via Auto-Regressive Context Modeling
Li, D., Dong, R., Li, K., Chen, Y., Ding, D., & Li, L. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-36846-1_20 · Cited by 5
FUMO: Prior-Modulated Diffusion for Single Image Reflection Removal
Xu, T., Zhang, C., Zhai, G., & Liu, X. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-37359-5_30 · Cited by 2
Singla, K., Pandey, R., & Ghanekar, U. (2026) · Circuits Systems and Signal Processing · doi:10.1007/s00034-026-03555-w
“…six widely recognized benchmark datasets: Set5 [1], Set14 [43], BSD100 [26], Urban100 [15], PIRM_val [2], and DIV2K_val [39] using three metrics, including PSNR (Peak Signal-to-Noise Ratio) [13], SSIM (Structural Similarity Index) [41], and LPIPS (Learned Perceptual Image Patch Similarity) [45].”
Nwankwo, I. S., Male, F., & Karpyn, Z. T. (2026) · Transport in Porous Media · doi:10.1007/s11242-026-02300-w
CT-to-MRI translation of medical volume data based on an enhanced diffusion model
Ma, J., Chen, J., & Liang, A. (2026) · Scientific Reports · doi:10.1038/s41598-026-45181-1
“We select three metrics: MS-SSIM3 37 , FSIM 38 and UQI 39 .”
A Multi-Dimensional Evaluation Framework for Stable Diffusion Models
Hong, S. (2025) · IEEE International Conference on Electrical, Automation and Computer Engineering (ICEACE) · doi:10.1109/iceace67491.2025.11439725
“Another commonly used metric is Multi-Scale Structural Similarity (MS-SSIM), which measures the similarity of generated samples by comparing similarity across multiple scales [8].”
Han, B. H., Xu, R., Peng, L., et al. (2025) · International Conference on Computer and Communications (ICCC) · doi:10.1109/iccc68654.2025.11437646
“2) Evaluation Metrics: The widely adopted per-pixel metric Peak Signal-to-Noise Ratio (PSNR) [13] is selected as the evaluation criterion.”
Knowledge-Base and Contrastive-Language-Image-Pretraining Assisted Semantic Image Transmission
Li, C., He, Y., Zhang, T., et al. (2025) · International Conference on Computer and Communications (ICCC) · doi:10.1109/iccc68654.2025.11438037
“Consequently, their performance is evaluated using traditional metrics such as Peak Signal-to-Noise Ratio (PSNR) or Multi-Scale Structural Similarity (MS-SSIM) [7].”
Learned Image Compression with Frequency Feature Interaction and Non-local Cross-similarity Prior
Wang, J., & Ling, Q. (2026) · ACM Transactions on Multimedia Computing Communications and Applications · doi:10.1145/3803542
“For the MSE-based evaluation, PSNR is computed as For the MS-SSIM[41] evaluation, the average MS-SSIM is transformed into decibels as To quantitatively compare the overall rate-distortion performance, we adopt the Bjontegaard delta rate (BD-rate) metric[5], which measures the average bitrate…”
VolDiT: Controllable Volumetric Medical Image Synthesis with Diffusion Transformers
Seyfarth, M., Dar, S. U. H., Frisch, Y., Wild, P., Frey, N., André, F., & Engelhardt, S. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-38189-7_57 · Cited by 1
Sun, Y., Abe, M., & Joshita, S. (2026) · Advanced Biomedical Engineering · doi:10.14326/abe.15.206 · Cited by 2
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.