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 4,101–4,200 of 6,047
Objective Prediction of Human Visual Acuity Using Image Quality Metrics
Espinosa, J., Rodríguez, J. P., Mas, D., Ferri, C. V., & Perales, E. (2023) · Applied Sciences · doi:10.3390/app13106350 · Cited by 2
Feature Attention as a Control Mechanism for the Balance of Speed and Accuracy in Visual Search
Griffith, T., Townend, F. J., Baker, S.-A., & Lepora, N. F. (2023) · Computational Brain & Behavior · doi:10.1007/s42113-023-00171-8 · Cited by 2
Image Structure-Induced Semantic Pyramid Network for Inpainting
Huang, R., & Zheng, Y. (2023) · Applied Sciences · doi:10.3390/app13137812 · Cited by 2
High Layer Coding for Video Transmission
Semenov, S., Hu, S., & Tian, K. (2023) · International Conference on Computer, Information and Telecommunication Systems · doi:10.1109/cits58301.2023.10188777 · Cited by 2
“Structural similarity (SSIM) metric [13] was used to estimate the quality of the received video.”
MethodFIANCEE: Faster Inference of Adversarial Networks via Conditional Early Exits
Karpikova, P., Radionova, E., Yaschenko, A., Andrei, S., Kostyushko, L., Fabbricatore, R., & Aleksei, I. (2023) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52729.2023.01158 · Cited by 6
“Additionally, we imposed feature matching [78], VGG19 perceptual [39], L1 and MS-SSIM [80] losses.”
MethodInfluentialMIXLIC: Mixing Global and Local Context Model for learned Image Compression
Ruan, H., Wang, F. R., Xu, T., et al. (2023) · IEEE International Conference on Multimedia and Expo (ICME) · doi:10.1109/icme55011.2023.00123 · Cited by 2
“We optimize our model with two metrics (MSE and MS-SSIM) under different quality levels.”
MethodInfluential3D reconstruction quality assessment using 2D reprojection with dynamic partitioning
Rivera, C. C., & Salazar-Jimenez, A. E. (2023) · The Imaging Science Journal · doi:10.1080/13682199.2023.2248445 · Cited by 3
Xu, C., Xu, J., Huang, T., Zhang, H., Mei, L., Zhang, X., Duan, Y., & Yang, W. (2023) · International Journal of Remote Sensing · doi:10.1080/01431161.2023.2255352 · Cited by 3
“We quantify the difference between the network prediction (network’s output) and the real label (network’s input) by using the multi-scale structure similarity loss (MS-SSIM).”
MethodInfluentialA Novel Hybrid Regularization Method for Solving Inverse Scattering Problems
Liu, Y., Zhu, Z., & Zhang, B. (2023) · IEEE Transactions on Antennas and Propagation · doi:10.1109/tap.2023.3323083 · Cited by 2
“In Section IV-B, we add the structural similarity (SSIM) [48] to evaluate the performance of all methods, which is defined as SSIM (cid:0) ¯¯ ε r , ¯¯ ε ∗ r (cid:1) = (cid:2) l (cid:0) ¯¯ ε r , ¯¯ ε ∗ r (cid:1)(cid”
Deformable Dynamic Sampling and Dynamic Predictable Mask Mining for Image Inpainting
Shang, C., Zeng, Y., Yang, S., Jia, X., Lu, H., & He, Y. (2023) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2023.3316123 · Cited by 2
“We use L1 loss, peak signal-to-noise ratio (PSNR) [44], structural similarity index (SSIM) [45], and FID [45] to compare the performance of different methods as they are the most commonly used metrics in image inpainting.”
Enhanced Pixel Privacy: Leveraging Deep Learning for Advanced Image Steganography
Kashif, M., Shakeel, I., Ahmad, S., et al. (2023) · International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (I-SMAC) · doi:10.1109/i-smac58438.2023.10290153 · Cited by 2
“[38]:”
Dehazing in hyperspectral images: the GRANHHADA database
Fernández-Carvelo, S., Martínez‐Domingo, M. Á., Valero, E. M., & Andrés, J. H. (2023) · Scientific Reports · doi:10.1038/s41598-023-46808-3 · Cited by 5
“(3), where W 1 , W 2 and W 3 are relative weights with values 0.5932, 0.3031 and 0.1037 respectively, for the metrics MSSSIM 42 , VIF 47 Since the weights W i have been calculated specifically for image subset 1, if the metric were to be used for different images or different distortions (for…”
MethodInfluentialUnderwater Image Enhancement of Nuclear Power Plant Based on U-Net Model
Wang, P., Zhang, S., Zhang, K., et al. (2023) · Proceedings of the Annual Conference of the IEEE Industrial Electronics Society · doi:10.1109/iecon51785.2023.10311980 · Cited by 2
“In our experiment, we set M =5 and α =0.84, according to [12,13].”
BackgroundMethodInfluentialManipulation of Age Variation Using StyleGAN Inversion and Fine-Tuning
Yoon, D., Kim, J., Lorant, V., & Kang, S. (2023) · IEEE Access · doi:10.1109/access.2023.3336401 · Cited by 2
“In Table 3, we observed that the proposed method outperformed the existing method across five commonly used metrics: MS-SSIM [54], PSNR, mean KID, FID [59], and LPIPS [53].”
MethodFull-Reference Image Quality Assessment Using Self-Attention and Multiscale Features
Li, Y., Liao, X., Zhou, M., Ji, C., Wei, X., & Yue, H. (2023) · Journal of Circuits Systems and Computers · doi:10.1142/s021812662450155x · Cited by 2
Validation of image stream hashing: A forensic method for content verification
Wales, G. S. (2023) · Journal of Forensic Sciences · doi:10.1111/1556-4029.15432 · Cited by 2
“A negative number between zero (0) and minus one (−1) indicates a negative relationship between the reference and the questioned image Mean multi-scale structural similarity [ 20] A perfect match of the multi-scale structural similarity between the reference and the questioned image equals one (1).”
BackgroundAuxiliary Information Guided Self-attention for Image Quality Assessment
Yang, J., Wang, Z., Wang, G., Huang, B., Yang, Y., & Tu, W. (2023) · ACM Transactions on Multimedia Computing Communications and Applications · doi:10.1145/3635716 · Cited by 5
“…defined (related to image quality) perceptual differences between the inputs, such as error-based visibility [4, 52], structural similarity [51, 54], image information content-based [46, 47], contrast visibility [33, 34], or various other HVS-inspired feature-based similarities [42, 64,…”
BackgroundSelf-citationComparison of Metrics for Predicting Image and Video Quality at Varying Viewing Distances
Hammou, D., Krasula, L., Bampis, C. G., Li, Z., & Mantiuk, R. (2023) · IEEE International Workshop on Multimedia Signal Processing (MMSP) · doi:10.1109/mmsp59012.2023.10337748 · Cited by 5
“MS-SSIM [22] FR-Image 57.09 (1.5) A multi-scale version of SSIM.”
InfluentialTexture Atlas Compression Based on Repeated Content Removal
Luo, Y., Jin, X., Pan, Z., et al. (2023) · SIGGRAPH Asia Conference Papers · doi:10.1145/3610548.3618150 · Cited by 2
“We evaluate the appearance quality of 3D models using the generated texture via PSNR and MS-SSIM [Wang et al. 2003], which is computed by sampling a set of 50 camera views from a sphere surface of radius 2.5 centered around the model that is resized into the box [− 1 , 1 ] 3 .”
MethodInfluentialDemystifying the QoS and QoE of Edge-hosted Video Streaming Applications in the Wild with SNESet
Li, Y., Deng, G., Bai, C., Yang, J., Wang, G., Zhang, H. H., Bai, J., Yuan, H., Xu, M., & Wang, S. (2023) · Proceedings of the ACM on Management of Data · doi:10.1145/3626723 · Cited by 3
“The multimedia community mainly focuses on visual quality assessment, traditionally relying on pixel-level patterns and data-driven models (e.g., PSNR [51], SSIM [108], and VMAF [80]) as well as deep learning models [32, 118, 122] to model users’ perception of encoded video.”
MethodFacial Image Reconstruction and its Influence to Face Recognition
Pleško, F., Goldmann, T., & Malinka, K. (2023) · International Conference of the Biometrics Special Interest Group (BIOSIG) · doi:10.1109/biosig58226.2023.10346000 · Cited by 2
“03 , L = 255 [16].”
Efficient, Self-Supervised Human Pose Estimation with Inductive Prior Tuning
Yoo, N., & Russakovsky, O. (2023) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw60793.2023.00351 · Cited by 2
“While earlier works measure perceptual against pixel-wise loss [11, 14, 32], recent works in reconstruction use pixel-wise L 1 or L 2 in conjunction with perceptual loss [30, 38, 44].”
MethodFan, C.-C., Yang, H., Peng, L., et al. (2024) · IEEE Transactions on Cognitive and Developmental Systems · doi:10.1109/tcds.2023.3345922 · Cited by 2
“We use the same metrics as TR-GAN to evaluate the performance of proposed model, i.e., MMSE, multiscale structural similarity metric (MS-SSIM) [41], and peak signal-to-noise ratio (PSNR).”
MethodEditableNeRF: Editing Topologically Varying Neural Radiance Fields by Key Points
Zheng, C., Lin, W., & Xu, F. (2024) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2024.3366148 · Cited by 2
Fourier Feature Activated Neural-field Modelling For Human Avatar Representation
Yang, Q., Weng, D., & Zhang, H. (2023) · IEEE Smart World Congress (SWC) · doi:10.1109/swc57546.2023.10449023 · Cited by 2
“PSNR and SSIM [9] were used to assess the signal percentage over noise and the similarity score between the predicted representation and the ground truth, respectively.”
MethodResultInfluentialReznik, Y. A. (2024) · doi:10.1145/3638036.3640295 · Cited by 2
Re-identification Robustness Over Time - The Case for Synthetic Training Data
Matzke, S., Pionzewski, C., Rutinowski, J., Çelik, F., & König, M. (2023) · International Conference on Machine Learning and Applications · doi:10.1109/icmla58977.2023.00206 · Cited by 2
“These include the well-known Mean Squared Error [19], the Structural Similarity Index Measure [20], and the Multi-Scale Structural Similarity Index Measure (MSSSIM) [21], which measures structural similarity at multiple scales.”
MethodDeep Conditional HDRI: Inverse Tone Mapping via Dual Encoder-Decoder Conditioning Method
Nam, Y., Kim, J., Shim, J.-H., & Kang, S. (2024) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2024.3379890 · Cited by 3
“For fair performance comparison on the generated HDR images, PU21 [43] encoding was applied and compared in terms of PSNR, SSIM [44], MS-SSIM [45], HDR-VDP-3 Quality (Q), and HDR-VDP-3 just objectionable differences (JOD) indicators, which improved HDR-VDP-2 [46], [47], defined in the range of 0–10,…”
MethodTIST-Net: style transfer in dynamic contrast enhanced MRI using spatial and temporal information
Tattersall, A. G., Goatman, K. A., Kershaw, L. E., Semple, S. I. K., & Dahdouh, S. (2024) · Physics in Medicine and Biology · doi:10.1088/1361-6560/ad4193 · Cited by 2
Semantics-Guided and Saliency-Focused Learning of Perceptual Video Compression
Li, B. (2024) · IEEE Access · doi:10.1109/access.2024.3396407 · Cited by 2
“We abstain from employing traditional PSNR and SSIM [60] metrics due to their inconsistency with the human visual experience.”
MethodContext-Aware Transformer GAN for Direct Generation of Attenuation and Scatter Corrected PET Data
Jafaritadi, M., Anaya, E., Chinn, G., Rosenberg, J., Liang, T., & Levin, C. S. (2024) · IEEE Transactions on Radiation and Plasma Medical Sciences · doi:10.1109/trpms.2024.3397318 · Cited by 3
“For quantitative evaluation of the global image quality, our study utilized peak signal-to-noise ratio (PSNR) [55], multiscale structural similarity index (MS-SSIM) [56], normalized mean-squared error (NRMSE), and MAE as global assessment metrics.”
MethodBoundary-aware residual network for defect detection in strip steel products
Hu, C., Ma, R., Du, X., & Gong, J. (2024) · Evolving Systems · doi:10.1007/s12530-024-09588-3 · Cited by 2
“This fusion loss consists of the Structural Similarity Index (SSIM), Binary Cross-Entropy (BCE), and Intersection over Union (IOU).”
InfluentialA visual simulation method for 3D displays with dense views and experimental verification
Zhang, S., Xing, S., Yu, X., et al. (2024) · Optics Communications · doi:10.1016/j.optcom.2024.130662 · Cited by 2
Compressed Point Cloud Quality Index by Combining Global Appearance and Local Details
Xu, Y., Zhang, Y., Yang, Q., Xu, X., & Liu, S. (2024) · ACM Transactions on Multimedia Computing Communications and Applications · doi:10.1145/3672567 · Cited by 4
Fast First Pass in Two-Pass Video Encoding Using Sub-Sampling
Henkel, A., Helmrich, C. R., Hinz, T., et al. (2024) · Picture Coding Symposium · doi:10.1109/pcs60826.2024.10566444 · Cited by 2
Temporal Enhanced Hybrid Neural Representation for Video Compression
Wang, J., Liu, Y., Zhu, S., et al. (2024) · Picture Coding Symposium · doi:10.1109/pcs60826.2024.10566352 · Cited by 2
“The L2-norm serves as the loss function, while evaluation metrics encompass PSNR, MS-SSIM [21], and Bits Per Pixel (bpp).”
MethodINNOVATIONS IN MODERN EDUCATION, THEIR APPLICATION TRENDS AS ONE OF THE MAIN ISSUES IN PEDAGOGY
Amiraslanova, T. M. T. A. T., & Ismayilova, R. I. R. (2024) · PIRETC-Proceeding of The International Research Education & Training Centre · doi:10.36962/piretc31022024-32 · Cited by 2
Using Spatial‐Temporal Attention for Video Quality Evaluation
Chi, B., Su, R., & Chen, X. (2024) · International Journal of Intelligent Systems · doi:10.1155/2024/5514627 · Cited by 2
A Spherical Light Field Database for Immersive Telecommunication and Telepresence Applications
Zerman, E., Gond, M., Takhtardeshir, S., et al. (2024) · International Conference on Quality of Multimedia Experience (QoMEX) · doi:10.1109/qomex61742.2024.10598264 · Cited by 2
“Structural Similarity Index (SSIM) [34] and its Multi-scale implementation MS-SSIM [35] rely on perceived changes in structural information, providing insights into human visual perception.”
MethodQuality Assessment of Light Field Images Based on Adaptive Attention in ViT
Du, Y., Lang, W., Hu, X., Yu, L., Zhang, H., Zhang, L., & Wu, Y. (2024) · Electronics · doi:10.3390/electronics13152985 · Cited by 3
Syntax-Guided Content-Adaptive Transform for Image Compression
Shi, Y., Ye, L., Wang, J., et al. (2024) · Sensors · doi:10.3390/s24165439 · Cited by 2
“When assessing image quality, this method not only employs the Peak Signal-to-Noise Ratio (PSNR) metric but also utilizes the Multi-Scale Structural Similarity (MS-SSIM) index [31].”
MethodKang, X., Wang, X., He, Y., Chen, B., & Xuelin, S. (2024) · IEEE International Conference on Multimedia and Expo (ICME) · doi:10.1109/icme57554.2024.10687355 · Cited by 2
“Improved versions of these metrics are compared with six state-of-the-art methods: MS-SSIM [30], FSIM [31], VIF [8], NLPD [4], MAD [25], PieAPP [32].”
MethodLi, J., Zheng, S., Zhu, C., Sun, Y., Chen, P., Shui, Z., Zhang, Y., Li, H., & Yang, L. (2024) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3664647.3681295 · Cited by 7
Gartner, C. M., Mathew, J. K., & Bullock, D. M. (2024) · Future Transportation · doi:10.3390/futuretransp4040062 · Cited by 3
“Several adaptations of the SSIM have also been developed to handle several different use cases: Wang’s Multiscale SSIM (MS-SSIM) [34] varies the viewing conditions (e.g., contrast); Sampat et al.’s Complex Wavelet SSIM (CW-SSIM) [19] has been shown to handle minute variations in the image better…”
Perceptual Neural Audio Coding With Modified Discrete Cosine Transform
Lim, H., Lee, J., Kim, B. H., Jang, I., & Kang, H.-G. (2024) · IEEE Journal of Selected Topics in Signal Processing · doi:10.1109/jstsp.2024.3491576 · Cited by 3
“…metrics such as mean squared error (MSE) and mean absolute error (MAE) on their relevance to the actual perceptual distortion have been discussed in several previousworks[4],[16],[46],[47],andvariouseffortshavebeen madetoinduceamoreaccuratereflectionofthehumanpercep-tionduringmodeltraining.”
Evaluation of REAL-ESRGAN Using Different Types of Image Degradation
Andriadi, K., Zarlis, M., Heryadi, Y., & Chowanda, A. (2024) · International Conference on ICT for Smart Society (ICISS) · doi:10.1109/iciss62896.2024.10751549 · Cited by 2
Image Manipulation Quality Assessment
Wu, X., Lou, J., Wu, Y., Liu, W., Rosin, P. L., Colombo, G. B., Allen, S. M., Whitaker, R. M., & Liu, H. (2024) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2024.3504854 · Cited by 3
“Popular traditional FR-IQA models include SIM [11], MS-SSIM [31], and VIF [22].”
Verma, S., Patil, J., & Tamhankar, I. (2024) · doi:10.1109/icspcre62303.2024.10675155 · Cited by 2
“…proposes a novel approach to this problem, using Conditional Generative Adversarial Networks (CGANs), designed and implemented through PyTorch for photorealistic image synthesis of the autonomous car scenario with some form of control and the ability to create details that surpass control [20].”
Li, M., Xu, Z., Li, Z., & Qi, Y. (2024) · International Journal of Computational Intelligence and Applications · doi:10.1142/s1469026824500329 · Cited by 2
Patch Inverter: A Novel Block-Wise GAN Inversion Method for Arbitrary Image Resolutions
Li, Y., Xu, M., Li, S., Zhang, J., & Guan, Z. (2024) · IEEE Signal Processing Letters · doi:10.1109/lsp.2024.3506859 · Cited by 2
Krawczyk, P., Gaertner, M., Jansche, A., Bernthaler, T., & Schneider, G. (2024) · Methods in microscopy · doi:10.1515/mim-2024-0012 · Cited by 4
Perceptual Quality Assessment for Pansharpened Images Based on Deep Feature Similarity Measure
Zhang, Z., Zhang, S., Meng, X., Chen, L., & Shao, F. (2024) · Remote Sensing · doi:10.3390/rs16244621 · Cited by 2
Investigating the effect of loss functions on single-image GAN performance
YILDIZ, E., Yüksel, M. E., & Sevgen, S. (2024) · Journal of Innovative Science and Engineering (JISE) · doi:10.38088/jise.1497968 · Cited by 2
Kageyama, H., Yoshida, N., Kondo, K., & Akai, H. (2024) · Radiological Physics and Technology · doi:10.1007/s12194-024-00871-1 · Cited by 3
A Novel Video Compression Approach Based on Two-Stage Learning
Shao, D., Wang, N., Chen, P., Liu, Y., & Lin, L. (2024) · Entropy · doi:10.3390/e26121110 · Cited by 5
“Mean squared error (MSE) [47] measured the discrepancy between the predicted values and the actual values.”
HNR-ISC: Hybrid Neural Representation for Image Set Compression
Zhang, P., Wang, S., Wang, M., Chen, P., Wu, W., Wang, X., & Kwong, S. (2024) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2024.3521715 · Cited by 8
“…measures: To evaluate the performance of our proposed approach, we employ Peak Signal-to-Noise Ratio (PSNR) and Multi-Scale Structural Similarity Index Measure (MS-SSIM) [47] as the evaluation measures because PSNR and MS-SSIM are commonly used as benchmarks in the field of image compression.”
Efficient Image Inpainting for Handwritten Text Removal Using CycleGAN Framework
Maiti, S., Panuganti, S. N., Bhatnagar, G., et al. (2025) · Mathematics · doi:10.3390/math13010176 · Cited by 2
Padding-enabled real-time high-fidelity temporal single pixel imaging
Keyaki, R., Matsuno, J., & Fukatsu, S. (2025) · Applied Physics Express · doi:10.35848/1882-0786/ada71b · Cited by 2
Adversarial learning for unguided single depth map completion of indoor scenes
Medhi, M., & Sahay, R. R. (2025) · Machine Vision and Applications · doi:10.1007/s00138-024-01652-x · Cited by 2
“We use multiscale SSIM (MS-SSIM) [ 97] to capture the perceptual differences between estimated and ground-truth depth maps.”
Synthetic Image Generation of Aortic Valves Using Conditional DDPM
Hofmann, M., Fromme, D., Streckert, T., & Thiem, J. (2024) · International Conference on Bioinformatics Research and Applications (ICBRA) · doi:10.1145/3700666.3700686 · Cited by 2
Scalable and High-Quality Neural Implicit Representation for 3D Reconstruction
Yang, L., Deng, B., & Zhang, J. (2025) · IEEE Transactions on Visualization and Computer Graphics · doi:10.1109/tvcg.2025.3530452 · Cited by 2
“Furthermore, wereportrenderingqualitymetrics—includingPeakSignalto NoiseRatio(PSNR),StructuralSimilarity(SSIM)[57],and LearnedPerceptualImagePatchSimilarity(LPIPS)[58]—both beforeandaftertheregistrationoptimizationtoillustratethe effectivenessofourapproach.”
Zein, H., Laurent, L., Fournier, R., & Naït-Ali, A. (2025) · Multimedia Tools and Applications · doi:10.1007/s11042-025-20639-y · Cited by 2
“Structural Similarity Structural similarity [18] is mainly used to assess the quality degradation of pictures after compression compared to the original one.”
Using deep feature distances for evaluating the perceptual quality of MR image reconstructions
Adamson, P. M., Desai, A., Dominic, J., Varma, M., Blüthgen, C., Wood, J. P., Syed, A., Boutin, R. D., Stevens, K. J., Vasanawala, S. S., et al. (2025) · Magnetic Resonance in Medicine · doi:10.1002/mrm.30437 · Cited by 13
Filling the Gaps: A Comprehensive Review of Image Inpainting
Bshibsh, T. S. A., & Abbadi, N. E. (2024) · Automation, Control, and Information Technology · doi:10.1109/acit62805.2024.10877030 · Cited by 2
RAIDS: Radar Range-Azimuth Map Estimation From Image, Depth, and Semantic Descriptions
Rangaraj, P. A., Alkanat, T., & Pandharipande, A. (2025) · IEEE Sensors Journal · doi:10.1109/jsen.2025.3540660 · Cited by 8
“The luminance, contrast, and structure, which are referred to as l , c , s , are defined as in [33].”
An “Opinion-Unaware” 3-D Image Quality Evaluator Using Correlation Features of Scene Components
Raghuwanshi, P. K., Poreddy, A. K. R., Appina, B., & Kokil, P. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2025.3542137 · Cited by 2
“D image QA (IQA) methods [4], [5], [6], [7] independently to each view and aggregate the individual scores for an overall score.”
Correlating Without Confidence: The Overlooked Role of Uncertainty When Ranking Objective Measures
Zizien, A., & Fliegel, K. (2025) · IEEE Access · doi:10.1109/access.2025.3544307 · Cited by 2
“Specifically, our evaluation includes results for CLIP-IQA+ [36], TOPIQ [37], hyperIQA [38], DB-CNN [39], TReS [40], WaDIQaM [41], PaQ-2-PiQ [42], BRISQUE [43], NIQE [44], MUSIQ [45], DISTS [46], PieAPP [47], LPIPS [48], VIF [49], FSIM [50], MS-SSIM [51], SSIM [52], and PSNR.”
A Biophysical-Based Skin Model for Heterogeneous Volume Rendering
Wang, Q., Luan, F., Dai, Y., Huo, Y., Bao, H., & Wang, R. (2025) · Computational Visual Media · doi:10.26599/cvm.2025.9450360 · Cited by 3
Li, N., Yang, X., Chen, T., Wang, T., & Ji, G. (2025) · Journal of Visual Communication and Image Representation · doi:10.1016/j.jvcir.2025.104417 · Cited by 2
LNeRV: Learnable Hierarchical Encoding Improve Neural Representation Video Codec
Chen, J., Liu, X., Chen, B., et al. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10888228 · Cited by 2
“Notably, although loss function of LNeRV is not optimized for SSIM, it performs nearly as well as or better than all NeRV-based methods in terms of MS-SSIM [35].”
Lightweight Image Quality Prediction Guided by Perceptual Ranking Feedback
Sekhri, A., Larabi, M., & Amirshahi, S. A. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10888173 · Cited by 2
“…[21], with our models achieving Methods CSIQ TID2013 KADID10K PLCC SROCC PLCC SROCC PLCC SROCC SSIM [8] 0.852 0.865 0.777 0.727 0.645 0.718 MS-SSIM [36] 0.889 0.906 0.830 0.786 0.686 0.802 VSI [37] 0.928 0.942 0.900 0.897 0.804 0.861 MAD [28] 0.950 0.947 0.827 0.781 0.586 0.724 GMSD [38] 0.945…”
InfluentialEnhancing Vision: Harmonizing Frequency for Imaging Quality and Perception Accuracy
Chen, H., & Ma, K. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10889903 · Cited by 2
“L s stands for Multi-scale Structural Similarity (MS-SSIM [35]) loss.”
Single-image dehazing method based on Rayleigh Scattering and adaptive color compensation
Guo, X., Sun, Q., Zhao, J., et al. (2025) · PLoS ONE · doi:10.1371/journal.pone.0315176 · Cited by 2
“The MS-SSIM is an algorithm used to compare the similarity between two images [33].”
Huang, Q., Lü, H., Liu, W., et al. (2025) · IEEE Sensors Journal · doi:10.1109/jsen.2025.3550525 · Cited by 2
“It should be noted that our model shows a significant improvement in MS-SSIM [36] in all datasets compared to VTM-13.”
InfluentialBlind HDR image quality assessment based on aggregating perception and inference features
Wan, D., Jiang, X., & Yu, Q. (2025) · Scientific Reports · doi:10.1038/s41598-025-94005-1 · Cited by 3
DocMatcher: Document Image Dewarping via Structural and Textual Line Matching
Hertlein, F., Naumann, A., & Sure-Vetter, Y. (2025) · IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) · doi:10.1109/wacv61041.2025.00563 · Cited by 2
“Typically, there are three visual evaluation metrics: Multiscale structural similarity (MS-SSIM) [32], Learned Perceptual Image Patch Similarity (LPIPS) [ 40], and, Local Distortion (LD) [36].”
Frequency-Aware Learned Image Compression Using Channel-Wise Attention and Restormer
Zhang, H., Jung, C., & Liu, X. (2025) · IEEE Access · doi:10.1109/access.2025.3559937 · Cited by 2
“We evaluate the proposed network by calculating the average RD performance in PSNR and MS-SSIM [30] on the Kodak24 and CLIC datasets [31], [32].”
A4RSIC: Attention-Augmented Adversarial Autoencoder for Remote Sensing Image Compression
Sivarajan, S., & Byju, A. P. (2025) · doi:10.1109/etis64005.2025.10961138 · Cited by 2
“Our results demonstrate that the A 4 RSIC using CBAM achieves significantly higher MS-SSIM [10] scores at lower bit rates compared to the other SOTA methods.”
User-Tailored Video Adaptation in Dynamic Environments
Choi, W., Chen, J., & Yoon, J. (2025) · IEEE Internet of Things Journal · doi:10.1109/jiot.2025.3564327 · Cited by 2
Image analysis approach based on VAE-GAN for IoT malware detection in urban environments
Dong, H., & Kotenko, I. V. (2025) · Innovations in Systems and Software Engineering · doi:10.1007/s11334-025-00610-8 · Cited by 2
A Multi-Level Patch Dataset for JPEG Image Quality Assessment by Absolute Binary Decision
Honda, S., Maeda, Y., Watanabe, O., & Fukushima, N. (2025) · IEEE Open Journal of Signal Processing · doi:10.1109/ojsp.2025.3571674 · Cited by 3
“These groups correspond to classical grayscale-based methods, perceptual metrics incorporating color and structure, and learning-based approaches: Y-PSNR, PSNR-HVS-M [32], SSIM [33], MS-SSIM [34], VIF [35], GMSD [36], MDSI [37], FSIM [38], VSI [39], SSIMU-LACRA2 [25], Butteraugli [40], VMAF (v0.”
Data-driven Reflectance Acquisition and Modeling for Predictive Rendering
Kavoosighafi, B. (2025) · Linköping studies in science and technology. Dissertations · doi:10.3384/9789181181494 · Cited by 2
Fourier convolutional decoder: reconstructing solar flare images via deep learning
Selcuk-Simsek, M., Massa, P., Xiao, H., Krucker, S., & Csillaghy, A. (2025) · Neural Computing and Applications · doi:10.1007/s00521-025-11283-6 · Cited by 2
“Details of the luminance , contrast , and structure functions can be found in [49].”
Window view clarity evaluation in electrochromic glazings
Pilechiha, P., Baghmahyari, N. G., & Barahimi, F. (2025) · Energy and Buildings · doi:10.1016/j.enbuild.2025.116040 · Cited by 2
Direct inversion for bidimensional distributions of aerosol characteristics
Sipkens, T. A., & Corbin, J. C. (2025) · Journal of Aerosol Science · doi:10.1016/j.jaerosci.2025.106636 · Cited by 3
Frequency-Aware Wavelet U-Net: A Hybrid Framework for Seismic Data High-Frequency Expansion
Su, Y., Song, J., Duan, W., et al. (2025) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2025.3582616 · Cited by 2
“The calculation methods for these related variables can be found in the paper by Wang et al. [25].”
Self-citationWavelet-Based Fusion for Image Steganography Using Deep Convolutional Neural Networks
Khalifa, A. S., & Yadav, Y. (2025) · Electronics · doi:10.3390/electronics14142758 · Cited by 2
Convolution-Friendly Image Compression with FHE
Mertens, A., Nicolas, G., & Rovira, S. (2025) · Lecture notes in computer science · doi:10.1007/978-3-031-97260-7_1 · Cited by 2
“The measure closest to how the human eye perceives images is the Structural Similarity Index (SSI) [32]: De(cid:28)nition 2 (Structural Similarity Index (SSI)) SSI is de(cid:28)ned as: ) where µ i is the pixel sample mean of image i , i is the variance of image i , σ 2 xy is the covariance of…”
Liu, Y., Zhang, T., & Huang, Y. (2025) · IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing · doi:10.1109/jstars.2025.3591848 · Cited by 4
“We use the Ms-SSIM [79] image quality assessment method to construct object-level and global-level loss functions.”
InfluentialEvaluating Image Quality Metrics as Loss Functions for Image Dehazing
Dobre-Baron, R., Savu-Jivanov, A., & Ancuți, C. (2025) · Sensors · doi:10.3390/s25154755 · Cited by 3
“Numerous variants of the PSNR and SSIM such as the multi-scale SSIM (MS-SSIM) [16] exist, as well as metrics employing various human priors, like PSNR-HVS [27] (based on the JPEG standard, itself derived from human perceptions of frequency [28]) or NLPD [29] (based on the human vs. digital range of…”
InfluentialJagpal, D., Chen, X., & Namboodiri, V. P. (2025) · 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/cvpr52734.2025.01698 · Cited by 4
“We also tested compatibility and performance on more advanced architectures, specifically SDXL [32] and SD3 Medium [8].”
HybridMQA: Exploring Geometry-Texture Interactions for Colored Mesh Quality Assessment
Sarvestani, A. S., Tang, S., & Wang, Z. (2025) · 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/cvpr52734.2025.01995 · Cited by 4
“[13] apply MSE and MS-SSIM [46] to 2D texture maps, combining these with geometric descriptors to evaluate colored meshes.”
Self-citationA Multi-Scale Spatial-Temporal Network for Wireless Video Transmission
Zhou, X., Huang, D., Qi, Z., et al. (2024) · IEEE Globecom Workshops (GC Wkshps) · doi:10.1109/gcwkshp64532.2024.11100995 · Cited by 2
“In this paper, we qualify the end-to-end video transmission performance of the proposed VDJSCC model using pixel-wise metric PSNR and the perceptual metric MS-SSIM [14].”
Laazoufi, A., Hassouni, M. E., & Cherifi, H. (2025) · Expert Systems with Applications · doi:10.1016/j.eswa.2025.129817 · Cited by 4
BPCLIP: A Bottom-up Image Quality Assessment from Distortion to Semantics Based on CLIP
Song, C., Chen, H., Zhāng, W., Zhu, H., Liu, S., Huang, H., & Jiang, F. (2025) · IEEE International Conference on Multimedia and Expo (ICME) · doi:10.1109/icme59968.2025.11209992 · Cited by 4
“Traditional methods like MS-SSIM [8], BRISQUE [9], and NIQE [10] downsample images to multiple scales and compute quality scores in parallel, but they often fail to derive meaningful quality representations from low-resolution images.”
Subjective Visual Quality Assessment for High-Fidelity Learning-Based Image Compression
Jenadeleh, M., Sneyers, J., Jia, P., et al. (2025) · International Conference on Quality of Multimedia Experience (QoMEX) · doi:10.1109/qomex65720.2025.11219943 · Cited by 2
Parida, B. K., Sunilkumar, A. P., Sen, A., et al. (2025) · IEEE Access · doi:10.1109/access.2025.3613789 · Cited by 2
“SSIM: The Structural Similarity Index Measure (SSIM) assesses image quality by focusing on structural information, luminance, and contrast, aiming to align with human visual perception [43].”
DvD: Unleashing a Generative Paradigm for Document Dewarping via Coordinates-based Diffusion Model
Zhang, W., Lu, H., Ning, M., Huang, X., Wang, W., Huang, K., & Wang, Q. (2025) · SIGGRAPH Asia Conference Papers · doi:10.1145/3757377.3763913 · Cited by 4
“Following most of previous models [Das et al. 2019; Ma et al. 2022, 2018], we adopt the MS-SSIM (Multi-scale Structural Similarity) [Wang et al. 2003], LD (Local Distortion) [You et al. 2017b], and AD (Aligned Distortion) [Ma et al. 2022] to evaluate differences between dewarped and flat ground…”
Performance analysis of full-reference objective image and video quality assessment metrics
Bondžulić, B. P., Pavlović, B., & Petrovic, V. S. (2018) · Vojnotehnicki glasnik · doi:10.5937/vojtehg66-12708 · Cited by 2
Preliminary filtering and lossy compression of noisy remote sensing images
Vozel, B., Zemliachenko, A., Abramov, S., et al. (2018) · doi:10.1117/12.2325102 · 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.