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 2,501–2,600 of 6,047
Inverse Modeling of Core Flood Experiments for Predictive Models of Sandstone and Carbonate Rocks
An, S., Wenck, N., Manoorkar, S., et al. (2023) · Water Resources Research · doi:10.1029/2023wr035526 · Cited by 9
“The structural similarity index from Wang et al. (2003) was also calculated for different scans to further adjust the core extraction locations before averaging the images yielded from repeated scans to reduce the noise.”
MethodZhang, W., Liu, Y., Chen, L., Shi, J., Hong, X., & Wang, X. (2024) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2024.3369654 · Cited by 9
“A diverse range of well-established metrics are used for the assessment of image quality, including PSNR (Peak Signal-to-Noise Ratio), MS-SSIM [68], SSIM [52], FSIM [69], NIQE [70], and LPIPS [71].”
MethodDepth Video Inter Coding Based on Deep Frame Generation
Li, G., Lei, J., Pan, Z., Peng, B., & Ling, N. (2024) · IEEE Transactions on Broadcasting · doi:10.1109/tbc.2024.3374103 · Cited by 10
“The objective quality of the synthesized intermediate view is measured by four types of metrics, which are WS-PSNR [49], IV-PSNR [50], MS-SSIM [51], and VMAF [52].”
MethodRevisiting Latent Space of GAN Inversion for Robust Real Image Editing
Katsumata, K., Vo, D. M., Liu, B., & Nakayama, H. (2024) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv57701.2024.00523 · Cited by 19
“The qualitative observations are also validated by the LPIPS, structural similarity index measure (SSIM), multi-scale SSIM (MS-SSIM) [43], and MSE.”
MethodIm, J. Y., Halliburton, S. S., Mei, K., Perkins, A. E., Wong, E., Roshkovan, L., Sandvold, O. F., Liu, L. P., Gang, G. J., & Noël, P. B. (2024) · Physics in Medicine and Biology · doi:10.1088/1361-6560/ad3dba · Cited by 9
Image Quality Assessment Tool for Conventional and Dynamic Magnetic Resonance Imaging Acquisitions
Nikiforaki, K., Karatzanis, I., Dovrou, A., Bobowicz, M., Gwoździewicz, K., Díaz, O., Tsiknakis, M., Fotiadis, D. I., Lekadir, K., & Kostas Marias, A. (2024) · Journal of Imaging · doi:10.3390/jimaging10050115 · Cited by 14
Open-Vocabulary Text-Driven Human Image Generation
Zhang, K., Sun, J., Sun, J., Zhang, K., Sun, Z., & Tan, T. (2024) · International Journal of Computer Vision · doi:10.1007/s11263-024-02079-7 · Cited by 14
“Multi-Scale Structural Similarity (MS-SSIM) evaluates the structural similarity between two images (Wang & Simoncelli, 2003).”
Transfer Deep Learning for Dielectric Profile Reconstruction in Microwave Medical Imaging
Xue, F., Guo, L., Bialkowski, A., & Abbosh, A. M. (2024) · IEEE Journal of Electromagnetics RF and Microwaves in Medicine and Biology · doi:10.1109/jerm.2024.3402048 · Cited by 9
Han, S., Huang, Y., Zhou, M., Wei, X., Jia, F., Zhuang, X., Cheng, F., Xiang, T., Feng, Y., Pu, H., et al. (2024) · IEEE Transactions on Broadcasting · doi:10.1109/tbc.2024.3410707 · Cited by 11
Content Adaptive Distributed Joint Source-Channel Coding for Image Transmission With Hyperprior
Li, Y., Chen, X., Deng, X., et al. (2024) · IEEE Transactions on Cognitive Communications and Networking · doi:10.1109/tccn.2024.3438371 · Cited by 9
Temporal Adaptive Learned Surveillance Video Compression
Zhao, Y. Q., Ye, M., Ji, L., Guo, H., & Zhu, C. (2024) · IEEE Transactions on Broadcasting · doi:10.1109/tbc.2024.3434736 · Cited by 9
“Perceptual visual quality, on the other hand, evaluates how the human eye perceives the quality of a video, utilizing methods like SSIM and Multiscale Structural Similarity (MS-SSIM) [74].”
Integrating Multitemporal SAR and Optical Information for Missing Optical Imagery Generation
Dong, C., Yang, G., Wang, Y., Sun, W., Meng, X., & Chen, B. (2024) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2024.3451558 · Cited by 20
Generative Adversarial Networks With Radiomics Supervision for Lung Lesion Generation
Li, J., Pan, S., Zhang, X., et al. (2024) · IEEE Transactions on Biomedical Engineering · doi:10.1109/tbme.2024.3451409 · Cited by 9
“To evaluate whether our network generates overfitted lesions, we compared every generated lesion with every real lesion in terms of the structural similarity index (SSIM) [38].”
HDCCT: Hybrid Densely Connected CNN and Transformer for Infrared and Visible Image Fusion
Li, X., He, H., & Shi, J. F. (2024) · Electronics · doi:10.3390/electronics13173470 · Cited by 10
ISP Meets Deep Learning: A Survey on Deep Learning Methods for Image Signal Processing
Santos, C. F. G. D., Arrais, R. R., Silva, J. V. S. D., Silva, M. H. M. D., Neto, W. B. G. D. A., Lopes, L. T., Bileki, G. A., Lima, I. O., Rondon, L. B., Souza, B. M. D., et al. (2025) · ACM Computing Surveys · doi:10.1145/3708516 · Cited by 15
“Finally, level one is trained with MSE, VGG, and SSIM loss [88] and performs corrections in the local color processing, noise removal, texture enhancement, etc.”
WAFI-VSC: Wireless Adaptive Frame Interpolation Video Semantic Communication
Qi, X., Ma, N., Bao, Z., et al. (2024) · International Conference on Wireless Communications and Signal Processing (WCSP) · doi:10.1109/wcsp62071.2024.10826690 · Cited by 9
Wan, Z., Yan, X., Li, Z., Fan, X., Zuo, W., & Zhao, D. (2025) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2025.3538511 · Cited by 11
“…method, we conducted performance comparisons with 14 image quality assessment (IQA) models, spanning four 2D-IQA methods (PSNR, SSIM [22], MS-SSIM [23], and BRISQUE [24]), three stereoscopic IQA (SIQA) methods (Chen [31], SINQ [32], and PAD-Net [13]), three omnidirectional IQA (OIQA) methods…”
A Review Toward Deep Learning for High Dynamic Range Reconstruction
Martins, G. D. L., Lopez-Cabrejos, J., Martins, J., Leher, Q. O., Ferreti, G. D. S., Carvalho, L. H. C., Lima, F. B., Paixão, T., & Alvarez, A. B. (2025) · Applied Sciences · doi:10.3390/app15105339 · Cited by 14
SVD Filter Based Multiscale Approach for Image Quality Assessment
Saha, A., Bhatnagar, G., & Wu, Q. M. J. (2012) · 2012 IEEE International Conference on Multimedia and Expo Workshops · doi:10.1109/icmew.2012.15 · Cited by 11
Objective video presentation QoE predictor for smart adaptive video streaming
Wang, Z., Zeng, K., Rehman, A., Yeganeh, H., & Wang, S. (2015) · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · doi:10.1117/12.2187740 · Cited by 10
Self-citationA Novel Image Representation Framework Based on Gaussian Model and Evolutionary Optimization
Jiao, L., Zhang, S., Li, L., Yang, S., Liu, F., Hao, H., & Dong, H. (2016) · IEEE Transactions on Evolutionary Computation · doi:10.1109/tevc.2016.2603526 · Cited by 9
“When we look at other evaluation metrics, GMEO-E and GMEO-L almost achieve the best performance in most cases, except that for image Barbara under sampling rate 0.50, the MSSIM value of SCS is the highest.”
MethodInfluentialSubjective quality assessment of animation images
Yue, G.-H., Hou, C., & Gu, K. (2017) · Visual Communications and Image Processing · doi:10.1109/vcip.2017.8305135 · Cited by 10
“The selected FR metrics are SSIM [9], MSSSIM [10], FSIM [11], GMSD [12], GMSM [12], MAD [13], ADD-GSIM [14], IGM [15], SQMS [16], IW SCC [17] and GSS [18].”
MethodFusion of infrared and visible images with propagation filtering
Xing, C., Wang, Z., & Dong, C. (2018) · Infrared Physics & Technology · doi:10.1016/j.infrared.2018.09.016 · Cited by 9
Self-citationComputational Light Field Generation Using Deep Nonparametric Bayesian Learning
Meng, N., Sun, X., So, H. K., & Lam, E. Y. M. (2019) · IEEE Access · doi:10.1109/access.2019.2900153 · Cited by 8
“Better SSIM performance indicates that the DNB model enables the synthesis of light field with a more similar structure to the ground truth.”
InfluentialEffective crowd counting using multi-resolution context and image quality assessment-guided training
Li, H., Kong, W., & Zhang, S. (2020) · Computer Vision and Image Understanding · doi:10.1016/j.cviu.2020.103065 · Cited by 9
Do not Treat Boundaries and Regions Differently: An Example on Heart Left Atrial Segmentation
Zhao, Z., Puybareau, É., Boutry, N., & Géraud, T. (2021) · Proceedings - International Conference on Pattern Recognition/Proceedings/International Conference on Pattern Recognition · doi:10.1109/icpr48806.2021.9412755 · Cited by 12
Novel Ransomware Hiding Model Using HEVC Steganography Approach
Almomani, I., Alkhayer, A., & El‐Shafai, W. (2021) · Computers, materials & continua/Computers, materials & continua (Print) · doi:10.32604/cmc.2022.018631 · Cited by 9
“It is devised to determine the visual quality of a video frame based on multiple scales [38].”
BackgroundSelf-Mutating Network for Domain Adaptive Segmentation of Aerial Images
Lee, K., Lee, H., & Hwang, J. Y. (2021) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.00698 · Cited by 10
“…optimization is slow to train, the loss function using structural similarity is applied to generate a similar image as follows: where sim is the function to measure the structural similarity [35] between two images, and H, W, and C indicate the height, width, and channel of an image, respectively.”
MethodLee, D. Y., Kim, J.-H., Ko, H., & Bovik, A. C. (2022) · IEEE Transactions on Image Processing · doi:10.1109/tip.2022.3173810 · Cited by 9
Self-citationInterpretable Poisson Optimization-Inspired Deep Network for Single-Photon Counting Image Denoising
Tian, X. S., He, F., Liu, R., & Ma, J. (2022) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2022.3228266 · Cited by 13
“The metrics for the quantitative analysis include the peak signal-to-noise ratio (PSNR), structure similarity index (SSIM), multiscale SSIM (MS-SSIM) [30], information fidelity criterion (IFC) [31], and visual information fidelity (VIF) criterion [32].”
MethodOn the Importance of Denoising when Learning to Compress Images
Brummer, B., & Vleeschouwer, C. D. (2023) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv56688.2023.00247 · Cited by 19
“The paired image noise dataset is analyzed prior to training such that the multi-scale structural similarity (MSSSIM) [24] score between each noisy crop and its clean ground-truth is stored in a file, and the training dataset can be initialized such that all training crops exceed a set quality threshold.”
BackgroundRavi, K. S., Nandakumar, G., Thomas, N., Lim, M., Qian, E., Jimeno, M. M., Poojar, P., Jin, Z., Quarterman, P., Srinivasan, G., et al. (2023) · Frontiers in Neuroimaging · doi:10.3389/fnimg.2023.1072759 · Cited by 11
“These were: median object-masked local SNR, Peak SNR (PSNR, dB), Multi-scale Structural Similarity Index [MS-SSIM, (Wang et al., 2003)], the variance of the Laplacian, referred to as var-Lap (Pech-Pacheco et al., 2000), and MR Value.”
MethodLossy Image Compression in a Preclinical Multimodal Imaging Study
Cunha, F. F., Blüml, V., Zopf, L. M., et al. (2023) · Journal of Imaging Informatics in Medicine · doi:10.1007/s10278-023-00800-5 · Cited by 9
“Figure 8 provides a compact view of the multi-scale structural similarity index measure maps (MS-SSIM) [39], computed at regions depicted in Figs.”
MethodLSTM-Based Video Quality Prediction Accounting for Temporal Distortions in Videoconferencing Calls
Mittag, G., Naderi, B., Gopal, V., & Cutler, R. (2023) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49357.2023.10095711 · Cited by 13
“The most commonly used video quality metrics are PSNR (Peak Signal to Noise Ratio), SSIM [3], MS-SSIM [4], and VMAF [5], which are computed on a frame basis and with subsequent averaging across all frame scores to give the overall quality.”
MethodInfluentialUsing HVS Dual-Pathway and Contrast Sensitivity to Blindly Assess Image Quality
Chen, F., Fu, H. C., Yu, H., & Chu, Y. (2023) · Sensors · doi:10.3390/s23104974 · Cited by 10
“Some IQA methods [25,26] are designed for extracting multi-scale features based on this characteristic, and the results show that using multi-scale features can effectively improve the algorithm’s performance.”
MethodTurhal, U., Günay, A., & Nabiyev, V. V. (2023) · The Visual Computer · doi:10.1007/s00371-023-02866-2 · Cited by 10
Video Waterdrop Removal via Spatio-Temporal Fusion in Driving Scenes
Wen, Q., Wu, Y., & Chen, Q. (2023) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra48891.2023.10161240 · Cited by 18
“2) Quantitative Evaluation: a) PSNR and MS-SSIM: Due to the lack of paired data of real driving scenes, we evaluate the different methods on the testing set of our synthetic dataset.”
MethodImproving Visual Representation Learning Through Perceptual Understanding
Tukra, S., Hoffman, F., & Chatfield, K. (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/cvpr52729.2023.01392 · Cited by 10
“over the IN1K validation set using the following quantitative measures: L1 error, Peak Signal to Noise Ration (PSNR), Structural Similarity Index (SSIM) [49], Inception Score (IS) [40] and Fréchet inception distance (FID) [19].”
MethodInfluentialTowards a Multiscale Point Cloud Structural Similarity Metric
Lazzarotto, D. N., & Ebrahimi, T. (2023) · IEEE International Workshop on Multimedia Signal Processing (MMSP) · doi:10.1109/mmsp59012.2023.10337634 · Cited by 9
“The correlation between image-based SSIM and human perception is usually higher when scores are computed at multiple scales and pooled together [9], even when considering artifacts created by learning-based codecs [10].”
BackgroundMethodYu, X., Pan, J. H., Xu, J., et al. (2024) · ISPRS Journal of Photogrammetry and Remote Sensing · doi:10.1016/j.isprsjprs.2024.01.007 · Cited by 9
RGB-T Saliency Detection Based on Multiscale Modal Reasoning Interaction
Wu, Y., Jia, T., Chang, X., Wang, H., & Chen, D. (2024) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2024.3419115 · Cited by 11
“The boundary loss has two components, a local structural similarity (SSIM) index [41] loss and a global edge loss.”
BackgroundAn End-to-End Real-World Camera Imaging Pipeline
Xu, K., Ma, Z., Xu, L., He, G., Li, Y., Yu, W., Han, T., & Yang, C. (2024) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3664647.3680805 · Cited by 15
“Inspired by the principles of intra-frame coding in traditional video compression, the CSA is designed to identify and leverage non-local redundancy within feature representations, promoting more efficient compression.”
On Perceptual Encryption: Variants of DCT Block Scrambling Scheme for JPEG Compressed Images
Khan, M. I., Jeoti, V., & Malik, A. S. (2010) · Communications in computer and information science · doi:10.1007/978-3-642-17641-8_26 · Cited by 9
No-reference quality assessment for DCT-based compressed image
Wang, C., Shen, M., & Yao, C. (2015) · Journal of Visual Communication and Image Representation · doi:10.1016/j.jvcir.2015.01.006 · Cited by 10
Energy Preserving QMF for Image Processing
Lian, J., & Wang, Y. (2014) · IEEE Transactions on Image Processing · doi:10.1109/tip.2014.2326772 · Cited by 9
Video Quality Assessment based on Adaptive Block-size Transform Just-Noticeable Difference model
Ma, L., Zhang, F., Li, S., & Ngan, K. N. (2010) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2010.5649188 · Cited by 9
“The most popular Image Quality Assessment (IQA) and VQA algorithms that embody this approach include Structural SIMilarity (SSIM) index [5]-[7], Visual Information Fidelity (VIF) [9], and Video Quality Metric (VQM) [8].”
MethodOn combining visual perception and color structure based image quality assessment
Lu, W. B., Xu, T., Ren, Y., et al. (2016) · Neurocomputing · doi:10.1016/j.neucom.2016.01.117 · Cited by 9
Multiscale-SSIM index based stereoscopic image quality assessment
Khan, S., & Channappayya, S. S. (2016) · National Conference on Communications · doi:10.1109/ncc.2016.7561082 · Cited by 9
“To assess the quality of luminance images we rely on the robust MS-SSIM [12] index.”
MethodQuality Assessment of Compressed MR Medical Images using General Regression Neural Network
Kumar, B., Sinha, G. R., & Thakur, K. (2011) · Cited by 10
A Human Perception Based Performance Evaluation of Image Quality Metrics
Wajid, R., Mansoor, A. B., & Pedersen, M. (2014) · Lecture notes in computer science · doi:10.1007/978-3-319-14249-4_29 · Cited by 11
“The VIF, IFC, SSIM and FSIMc performance for different distortion types is depicted in Figure 3.”
MethodInfluentialA perceptually relevant shearlet-based adaptation of the PSNR
Bosse, S., Siekmann, M., Samek, W., & Wiegand, T. (2017) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2017.8296294 · Cited by 10
“With β being selected distortion specific, the proposed method shows superior performance compared to PSNR, SSIM and MS-SSIM.”
BackgroundMethodInfluentialYang, Z., Yang, Z., & Gui, G. (2018) · Sensors · doi:10.3390/s18041175 · Cited by 9
“…four objective image quality metrics—the peak signal noise ratio (PSNR) in dB, the measure of structural similarity index (SSIM) [40], the multiscale SSIM (MS-SSIM) [41], and the feature similarity index (FSIM) [42]—are used to measure the performance of the proposed models for the test images.”
MethodBosq, T. W. D., Agarwal, S., Dijk, J., et al. (2018) · doi:10.1117/12.2307852 · Cited by 9
relf: robust regression extended with ensemble loss function
Hajiabadi, H., Monsefi, R., & Yazdi, H. S. (2018) · Applied Intelligence · doi:10.1007/s10489-018-1341-9 · Cited by 11
“functions lead to different Optimum Bayes with possibly different characteristics. Thus, in each environment the choice of the underlying loss function is important, as it will impact the performance[44,48]. Letting ^ denote the estimated parameter of a correct parameter , the loss function L(^;) is a positive functio…”
BackgroundRectangular-Normalized Superpixel Entropy Index for Image Quality Assessment
Lü, T., Wang, J., Zhou, H., Jiang, J., Ma, J., & Wang, Z. (2018) · Entropy · doi:10.3390/e20120947 · Cited by 9
“The evaluation results are compared with some representative FR IQA metrics, including some state-of-art algorithms: • SSIM [14]—a widely used methods for IQA • MS-SSIM [15]—the multi-scale SSIM (MS-SSIM) of wavelet domain • FSIM [17]—the best-performance IQA method based on structural information • VSNR [12]—wavelet-…”
MethodInfluentialSelf-citationMachine vision quality assessment for robust face detection
Soundararajan, R., & Biswas, S. (2018) · Signal Processing Image Communication · doi:10.1016/j.image.2018.12.012 · Cited by 10
“similarity index [20] is used the predict the performance of object”
MethodQuality Metric Aggregation for HDR/WCG Images
Rousselot, M., Ducloux, X., Meur, O. L., & Cozot, R. (2019) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2019.8803635 · Cited by 10
“These features are inspired by the MS-SSIM features [7], the multi-scale version of the SSIM index [6].”
MethodA no-reference framework for evaluating video quality streamed through wireless network
Uzair, M., Dony, R. D., Jamil, M., Khawaja, B. A., & Khan, M. N. (2019) · TURKISH JOURNAL OF ELECTRICAL ENGINEERING & COMPUTER SCIENCES · doi:10.3906/elk-1811-173 · Cited by 10
“an MSSIM value is obtained for each of the P and B frames by taking the I frame as reference, and P and B frames as distorted [44].”
MethodInfluentialLearning No-Reference Quality Assessment of Multiply and Singly Distorted Images With Big Data
Zhang, Y., Mou, X., & Chandler, D. M. (2019) · IEEE Transactions on Image Processing · doi:10.1109/tip.2019.2952010 · Cited by 9
“The six FR methods were SSIM [34], multiscale structure similarity (MS-SSIM) [39], feature similarity index (FSIM) [40], visual information fidelity (VIF) [21], most apparent distortion (MAD) [52], and gradient magnitude similarity deviation (GMSD) [32].”
MethodAerial image classification by learning quality-aware spatial pyramid model
Huang, F., Qi, X., Li, C., & Hu, W. (2020) · Future Generation Computer Systems · doi:10.1016/j.future.2020.04.047 · Cited by 11
Blind First-Order Perspective Distortion Correction Using Parallel Convolutional Neural Networks
Gallego, N. P. D., Ilao, J. P., & Cordel, M. O. (2020) · Sensors · doi:10.3390/s20174898 · Cited by 9
“Model C appears to have the lowest transformation matrix error among other variants but the lowest pixel-wise error and highest SSIM accuracy of the corrected images were produced by Model A.”
MethodMulti-Resolution Aitchison Geometry Image Denoising for Low-Light Photography
Miller, S., Zhang, C., & Hirakawa, K. (2021) · IEEE Transactions on Image Processing · doi:10.1109/tip.2021.3087943 · Cited by 10
“In addition, we adopt multiscale SSIM to compute the reference and recovered Aitchison variables at each subband [38]—we report the MS-SSIM score, averaged over the various Aitchison subbands.”
MethodAppearance-preserving tactile optimization
Tymms, C., Wang, S., & Zorin, D. (2020) · ACM Transactions on Graphics · doi:10.1145/3414685.3417857 · Cited by 10
“Well-established visual metrics include those based on structural similarity : SSIM [Wang et al. 2004], FSIM [Zhang et al. 2011], MSSIM [Wang et al. 2003].”
MethodEnd-to-End Image Patch Quality Assessment for Image/Video With Compression Artifacts
Pham, T. T., Hoang, X. V., Nguyen, N. T., Dinh, D. T., & Ha, L. T. (2020) · IEEE Access · doi:10.1109/access.2020.3040416 · Cited by 9
“The SSIM is not only developed into MS-SSIM metric [34], but also be applied for FR IQA, such as the FSIM [4], the SRSIM [35] or RFSIM [5] obtaining promising results.”
MethodSi, J., Yang, H., Huang, B., et al. (2021) · IET Image Processing · doi:10.1049/ipr2.12132 · Cited by 9
“Ma’s method and Liu’s method are based on conventional algorithm SSIM and MS-SSIM.”
MethodInfluentialAn extended context-based entropy hybrid modeling for image compression
Fu, H., Liang, F., Lei, B., Zhang, Q., Liang, J., Tu, C., & Zhang, G. (2021) · Signal Processing Image Communication · doi:10.1016/j.image.2021.116244 · Cited by 12
Patch Decoder-Side Depth Estimation In Mpeg Immersive Video
Milovanović, M., Henry, F., Cagnazzo, M., et al. (2021) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp39728.2021.9414056 · Cited by 9
“Moreover, we achieve 9.3% VMAF [13] BD-Rate reduction for medium bitrate, and 13.4% for low bitrate, while having 6.7% perceptual MS-SSIM [14] gain for medium bi-trate, with 11.6% gain for low bitrate.”
MethodInfluentialAn end-to-end sea fog removal network using multiple scattering model
An, S., Huang, X., Zheng, Z., et al. (2021) · PLoS ONE · doi:10.1371/journal.pone.0251337 · Cited by 9
“In terms of evaluation indexes [40, 41], PSNR [30] and SSIM [31] were used for quantitative evaluation.”
MethodHamilton, S. J., Isaacson, D., Kolehmainen, V., et al. (2021) · Inverse Problems and Imaging · doi:10.3934/ipi.2021032 · Cited by 9
“• Multi-scale Structural Similarity Index (MS-SSIM) - This is computed using Matlab’s multissim3, based on [65].”
MethodSelection and evaluation of spherical acquisition trajectories for industrial computed tomography
Bauer, F., Goldammer, M., & Grosse, C. (2021) · Proceedings of the Royal Society A · doi:10.1098/rspa.2021.0192 · Cited by 10
“remarked that these approaches have become too complicated to be applicable with real-world applications, in particular for the optimization of algorithms [82].”
BackgroundMultiple Task-Oriented Encoders for Unified Image Fusion
Li, Z., Liu, J., Liu, R., Fan, X., Luo, Z., & Gao, W. (2021) · IEEE International Conference on Multimedia and Expo (ICME) · doi:10.1109/icme51207.2021.9428212 · Cited by 11
“Correlation Coefficient (CC), Multi-Scale SSIM (MS-SSIM) [18] are employed to measure the correlation, structure similarity between source images and fused image.”
MethodMimicking Individual Media Quality Perception with Neural Network based Artificial Observers
Tiotsop, L. F., Mizdos, T., Barkowsky, M., Počta, P., Servetti, A., & Masala, E. (2022) · ACM Transactions on Multimedia Computing Communications and Applications · doi:10.1145/3464393 · Cited by 11
“In this work, we consider, as features, five video quality metrics, i.e. the PSNR[50], the SSIM[57] the MSSSIM[47], the VIF[41], and the VMAF[32], as well as six perceptual features, i.e. “Blockiness”, “Blockloss”, “Blur”, “Contrast”, “Flickering”, and “Noise”, which attempt to quantify how much each of the listed art…”
MethodPerceptual Image Compression using Relativistic Average Least Squares GANs
Cheng, Z., Fu, T., Hu, J., Guo, L., Wang, S., Zhao, X., Zhou, D., & Song, Y. (2021) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw53098.2021.00213 · Cited by 14
“Our submitted method ANTxNN SSIM is optimized by only MS-SSIM quality metric [36], i.e.”
MethodDetection of exact and near duplicates in phased-array ultrasound weld scan
Provençal, E., & Laperrière, L. (2021) · Procedia Manufacturing · doi:10.1016/j.promfg.2021.07.041 · Cited by 9
Progressive compressed records: taking a byte out of deep learning data
Kuchnik, M., Amvrosiadis, G., & Smith, V. (2021) · Proceedings of the VLDB Endowment · doi:10.14778/3476249.3476308 · Cited by 12
“Scan groups of 5 or higher have an MSSIM of 95%+, which is likely why they consistently perform well.”
MethodInfluentialOnline-trained Upsampler for Deep Low Complexity Video Compression
Klopp, J. P., Liu, K.-C., Chien, S., et al. (2021) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.00783 · Cited by 9
“Bjontegaard Deltas for Rate and Distortion as measured by PSNR and MS-SSIM over x265 with tunes psnr and ssim, respectively.”
MethodInfluentialMa, Y., Zhai, Y., Yang, J., et al. (2021) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3474085.3475671 · Cited by 9
GUV-Net for high fidelity shoeprint generation
Hassan, M. A., Wang, Y., Pang, W., Wang, D., Li, D., Zhou, Y., & Xu, D. H. (2021) · Complex & Intelligent Systems · doi:10.1007/s40747-021-00558-9 · Cited by 10
“To better assess the quality of SR images, we also used another highly adapted evaluation metric (MS-SSIM) based on the assumption of human visual system [94].”
MethodKang, M., Kim, H., Lee, S., & Han, S. (2021) · KSII Transactions on Internet and Information Systems · doi:10.3837/tiis.2021.11.013 · Cited by 9
“Therefore, a structural similarity measurement can be used as an approximate measure of similarity between images [32].”
MethodAnalysis of noisy image lossy compression by BPG using visual quality metrics
Коваленко, Б., Лукин, В. В., Науменко, В. В., et al. (2021) · IEEE International Conference on Advanced Trends in Information Theory (ATIT) · doi:10.1109/atit54053.2021.9678575 · Cited by 9
Image Quality Evaluation of Light Field Image Based on Macro-Pixels and Focus Stack
Meng, C., An, P., Huang, X., Yang, C., & Chen, Y. (2022) · Frontiers in Computational Neuroscience · doi:10.3389/fncom.2021.768021 · Cited by 8
“2D FR IQA metrics include PSNR, SSIM (Wang et al., 2004), multi-scale SSIM (MS-SSIM) (Wang et al., 2003), information weighting SSIM (IW-SSIM) (Wang and Li, 2010), FSIM (Zhang et al., 2011), FSIM based on Riesz transforms (RFSIM) (Zhang et al., 2010), noise quality measure (NQM) (Damera-Venkata et…”
MethodElimination of Defects in Mammograms Caused by a Malfunction of the Device Matrix
Tumakov, D. N., Kayumov, Z., Zhumaniezov, A., et al. (2022) · Journal of Imaging · doi:10.3390/jimaging8050128 · Cited by 9
“The following reference metrics are used here to assess the overlay efficiency: SSIM [38], VIF [39], VSI [40], CW_SSIM [41], MS_SSIM [42], FSIM [43], DISTS [44], GMSD [45], LPIPSVGG [46], and NLPD [47].”
MethodPerceptually Optimized Loss Function for Image Super-Resolution
Arezoomand, A., Cheraaqee, P., & Mansouri, A. (2021) · International Conference on Signal Processing and Intelligent Systems (ICSPIS) · doi:10.1109/icspis54653.2021.9729334 · Cited by 9
“Since MSE does not well correlate with human visual system (HVS) [20], [21], the super-resolved images are not perceptually plausible.”
BackgroundCIEGAN: A Deep Learning Tool for Cell Image Enhancement
Sun, Q., Yang, X., Guo, J., Zhao, Y., & Liu, Y. (2022) · Frontiers in Genetics · doi:10.3389/fgene.2022.913372 · Cited by 9
“The normalized root mean square error (NRMSE), peak signal-to-noise ratio (PSNR), and multi-scale structural similarity index (MS-SSIM) (Wang et al., 2003) are used in the image similarity assessment between the generated image and ground truth.”
MethodA pixel-wise framework based on convolutional neural network for surface defect detection
Dong, G. (2022) · Mathematical Biosciences & Engineering · doi:10.3934/mbe.2022408 · Cited by 9
Image Quality Evaluation in Professional HDR/WCG Production Questions the Need for HDR Metrics
Sugito, Y., Vázquez-Corral, J., Canham, T., et al. (2022) · IEEE Transactions on Image Processing · doi:10.1109/tip.2022.3190706 · Cited by 9
Two-Level Consistency Metric for Infrared and Visible Image Fusion
Lin, X., Zhou, G., Tu, X., Huang, Y., & Ding, X. (2022) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2022.3203115 · Cited by 9
Zhao, M., Lu, Z., Zhu, S., Wang, X., & Feng, J. (2022) · Medical Physics · doi:10.1002/mp.15988 · Cited by 9
Gram-GAN: Image Super-Resolution Based on Gram Matrix and Discriminator Perceptual Loss
Song, J. G., Yi, H., Xu, W., et al. (2023) · Sensors · doi:10.3390/s23042098 · Cited by 10
“Thereafter, a large number of PSNR-oriented methods [2–7] have emerged, which have the uniform property of a loss function consisting of a single mean square error (MSE), pixel-wise loss [8,9].”
MethodMeasuring and Predicting Perceptions of Video Quality Across Screen Sizes with Crowdsourcing
Bampis, C. G., Krasula, L., Li, Z., & Akhtar, O. (2023) · International Conference on Quality of Multimedia Experience (QoMEX) · doi:10.1109/qomex58391.2023.10178501 · Cited by 13
Möller, B., Pirklbauer, J., Klingner, M., Kasten, P., Etzkorn, M., Seifert, T. J., Schlickum, U., & Fingscheidt, T. (2023) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw59228.2023.00449 · Cited by 9
Multiprior Learning Via Neural Architecture Search for Blind Face Restoration
Yu, Y., Zhang, P., Zhang, K., Luo, W., & Li, C. (2023) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2023.3339614 · Cited by 16
“Following [9], [10], and [11], we employ widely used pixelwise metrics peak signal-to-noise ratio (PSNR) [70], structural similarity index (SSIM) [70], and multi-scale SSIM (MS-SSIM) [71] to evaluate the performance on CelebA-HQ with ground truth.”
A blind watermarking scheme using bitplane of DC component for JPEG Compressed images
Khan, M. I., & Jeoti, V. (2010) · International Conference on Emerging Technologies · doi:10.1109/icet.2010.5638498 · Cited by 9
Geometrical and statistical properties of vision models obtained via maximum differentiation
Malo, J., & Simoncelli, E. P. (2015) · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · doi:10.1117/12.2085653 · Cited by 15
Self-citationComparison of image quality assessment algorithms on compressed images
Charrier, C., Knoblauch, K., Moorthy, A. K., et al. (2010) · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · doi:10.1117/12.840221 · Cited by 9
Self-citationPerformance-analysis-based acceleration of image quality assessment
Phan, T. D., Sohoni, S., Chandler, D., & Larson, E. (2012) · 2012 IEEE Southwest Symposium on Image Analysis and Interpretation · doi:10.1109/ssiai.2012.6202458 · Cited by 10
“Indeed, modern image QA algorithms such as VIF [1], MS-SSIM [2], and MAD [3] are quite effective at QA, but they require a relatively large run-time— on the order of seconds.”
MethodHEVC enhancement using content-based local QP selection
Zhang, F., & Bull, D. (2016) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2016.7533154 · Cited by 9
Joskowicz, J., Sotelo, R., Garella, J. P., Zinemanas, P., & Simón, M. (2016) · IEEE Transactions on Broadcasting · doi:10.1109/tbc.2016.2617289 · Cited by 9
Combined full-reference image visual quality metrics
Ieremeiev, O., Lukin, V., Ponomarenko, N., Egiazarian, K., & Astola, J. (2016) · International Conference on Image Processing, Applications and Systems · doi:10.2352/issn.2470-1173.2016.15.ipas-180 · Cited by 14
“These metrics are the following: FSIM and its color version FSIMc [4], MSSIM [18], SSIM [19], VSNR [20], VIF and VIFP [21], NQM [22], WSNR [23], UQI [24], PSNR-HVS [25] and PSNR-HVS-M [26], PSNR-HMA and PSNR-HA [3], SFF [27], SRSIM [28], IWSSIM [29], IWPSNR [29], and MAD index [30].”
Efficient feature selection for Blind Image Quality Assessment based on natural scene statistics
Nizami, I. F., Majid, M., & Khurshid, K. (2017) · International Bhurban Conference on Applied Sciences and Technology (IBCAST) · doi:10.1109/ibcast.2017.7868071 · Cited by 10
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.