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 2,201–2,300 of 6,047
Park, W., & Kim, M. (2020) · IEEE Access · doi:10.1109/access.2020.3046040 · Cited by 14
“The multi-scale structural similarity index (MS-SSIM) [44] for an RGB color space, which is known to as a better metric for subjective image quality than PSNR, are used to measure the distortions in our experiments.”
MethodGradient Compass-Based Adaptive Multimodal Medical Image Fusion
Irshad, M. T., & Rehman, H. U. (2021) · IEEE Access · doi:10.1109/access.2021.3054843 · Cited by 12
“This index is used to measures the structural similarity between two images [25], in this case the similarity is between the source medical images and the fused medical image.”
MethodPant, S., Nooralishahi, P., Avdelidis, N. P., Ibarra‐Castanedo, C., Genest, M., Deane, S., Valdés, J. J., Zolotas, A. C., & Maldague, X. (2021) · Sensors · doi:10.3390/s21051604 · Cited by 13
“…and structural information between two images and compares them at different scales, where in each additional scale the images are passed through a low pass filter and down-sampled by 2 from the previous scale, providing a more advanced image quality measure [31] as compared to PNSR and ITF.”
MethodAdvanced Image Enhancement Method for Distant Vessels and Structures in Capsule Endoscopy
Rukundo, O., Pedersen, M., & Hovde, Ø. (2017) · NORA - Norwegian Open Research Archives · Cited by 12
Edge-competing Pathological Liver Vessel Segmentation with Limited Labels
Feng, Z., Wang, Z., Wang, X., Zhang, X., Cheng, L., Lei, J., Wang, Y., & Song, M. (2021) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v35i2.16221 · Cited by 13
“Qin et al. (2019) adopted the patch-level SSIM loss (Wang, Simoncelli, and Bovik 2003) to assign higher weights to the edges.”
MethodSelf-citationFoveated streaming of real-time graphics
Illahi, G. K., Siekkinen, M., Kämäräinen, T., et al. (2021) · ACM Multimedia Systems Conference (MMSys) · doi:10.1145/3458305.3463383 · Cited by 12
A brightness-adaptive kernel prediction network for inverse tone mapping
Cao, G., Zhou, F., Liu, K., et al. (2021) · Neurocomputing · doi:10.1016/j.neucom.2021.08.057 · Cited by 12
Dong, S., Hangel, G. J., Bogner, W., et al. (2021) · Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society · doi:10.1109/embc46164.2021.9630146 · Cited by 12
“M is the number of scale levels and the rest of the parameters were set to constant values suggested in [24].”
BackgroundMethodDisentangle Saliency Detection into Cascaded Detail Modeling and Body Filling
Song, Y., Tang, H., Sebe, N., & Wang, W. (2022) · ACM Transactions on Multimedia Computing Communications and Applications · doi:10.1145/3513134 · Cited by 16
“�SSIM improves the mask by enriching the representation with the structural information, while �IoU and �F helps the network to concentrate on the body and complements the body and detail map by ensuring the fused mask has a high F-score, respectively.”
InfluentialJPD-SE: High-Level Semantics for Joint Perception-Distortion Enhancement in Image Compression
Duan, S., Chen, H., & Gu, J. (2022) · IEEE Transactions on Image Processing · doi:10.1109/tip.2022.3180208 · Cited by 16
“achieve favorable rate-perception-accuracy-distortion performance as quantified via a large-scale user study, accuracy of a downstream bounding-box object detector, and three distortion measures (LPIPS [29], MS-SSIM [24], and PSNR).”
BackgroundMethodInfluentialNai, Y., Loi, H. Y., O’Doherty, S., Tan, T. H., & Reilhac, A. (2022) · Japanese Journal of Radiology · doi:10.1007/s11604-022-01311-z · Cited by 12
Sümbül, G., Xiang, J., & Demir, B. (2022) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2022.3204914 · Cited by 21
“…marginal distribution of quantized image latents d is the distortion metric, for which we utilize multiscale structural similarity index (MS-SSIM) [39] as It is worth noting that the proposed SCI-CBIR approach is independent of the image compression method utilized in the first step of our…”
MethodAhmad, I., & Shin, S. (2022) · Sensors · doi:10.3390/s22208074 · Cited by 11
“The y-axis is the recovered image quality represented as a multiscale structural similarity index measure (MS–SSIM) n dB against the compression savings given as the bitrate on the x–axis.”
MethodInfluentialPart-aware attention correctness for video salient object detection
Liu, Z.-Y., & Liu, J. (2022) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2022.105733 · Cited by 13
A Full- Reference Quality Assessment Metric for Cartoon Images
Li, C., Zhang, Z., Sun, W., Min, X., & Zhai, G. (2022) · IEEE International Workshop on Multimedia Signal Processing (MMSP) · doi:10.1109/mmsp55362.2022.9949270 · Cited by 14
“In C-IQA tasks, to prove the superiority of our metric, we choose state-of-the-art FR IQA metrics to compare with, including CW-SSIM [26], FSIM [27], GMSD [28], GSM [29], IFC [30], IW-SSIM [31], MS-SSIM [32], SFF [33], SFSN [7], SR-SIM [34], SSIM [4], UQI [35], VIF [5], VSI [6], VSNR [36].”
MethodExtremely Low-Light Image Enhancement with Scene Text Restoration
Hsu, P.-H., Lin, C.-T., Ng, C. C., Kew, J. L., Tan, M. Y., Lai, S., Chan, C. S., & Zach, C. (2022) · Proceedings - International Conference on Pattern Recognition/Proceedings/International Conference on Pattern Recognition · doi:10.1109/icpr56361.2022.9956716 · Cited by 15
“Inspired by [29], we adopted the multi-scale SSIM loss function in our work to enforce the image structure of the enhanced image I ′ to be close to that of the ground-truth image I : LSSIMMS = 1− SSIMMS(I ′, I ).”
MethodThe Impact of Black Edge Artifact on QoE of the FOV-Based Cloud VR Services
Song, J., Mao, X., & Yang, F. (2022) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2022.3232229 · Cited by 12
Li, B., Yang, T., & Zhao, X. (2023) · Journal of Applied Clinical Medical Physics · doi:10.1002/acm2.13908 · Cited by 12
Insights From Generative Modeling for Neural Video Compression
Yang, R., Yang, Y., Marino, J., & Mandt, S. (2023) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2023.3260684 · Cited by 15
“We note, however, other kinds of distortion functions such as MS-SSIM are possible, resulting in a different form of p(xt|z≤t) parameterized by x̂t [32].”
Methodψ-Net: Point Structural Information Network for No-Reference Point Cloud Quality Assessment
Xiong, J., Wu, S., Luo, W., et al. (2023) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49357.2023.10096258 · Cited by 16
Investigating the Visual Utility of Differentially Private Scatterplots
Panavas, L., Crnovrsanin, T., Adams, J. L., Ullman, J., Sarvghad, A., Tory, M. K., & Dunne, C. (2023) · IEEE Transactions on Visualization and Computer Graphics · doi:10.1109/tvcg.2023.3292391 · Cited by 17
“From our findings, we found that MS-SSIM [60], was the best way to predict the visual utility of a private plot.”
BackgroundMethodHuman Eye Contrast Sensitivity to Vehicle Displays under Strong Ambient Light
Qian, Y., Yang, Z., Hsiang, E., et al. (2023) · Crystals · doi:10.3390/cryst13091384 · Cited by 12
Axial super-resolution optical coherence tomography via complex-valued network
Wang, L., Chen, S., Liu, L., et al. (2023) · Physics in Medicine and Biology · doi:10.1088/1361-6560/ad0997 · Cited by 12
A Comprehensive Augmentation Framework for Anomaly Detection
Jiang, L., & Yan, Y. (2024) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v38i8.28720 · Cited by 22
Adaptive Super-Resolution Networks for Single-Pixel Imaging at Ultra-Low Sampling Rates
Liu, Z., Zhang, H., Zhou, M., Jiao, S., Zhang, X.-P. S., & Geng, Z. (2024) · IEEE Access · doi:10.1109/access.2024.3402693 · Cited by 14
“This is because they are defined based on pixel-level image differences [42], [43], [44].”
BackgroundNetwork Performance QoS Prediction
Frnda, J., Vozn̆ák, M., & Sevcik, L. (2014) · Advances in intelligent systems and computing · doi:10.1007/978-3-319-07776-5_18 · Cited by 12
Gradient-weighted structural similarity for image quality assessments
Li, Q., Fang, Y., Lin, W., & Thalmann, D. (2015) · International Symposium on Circuits and Systems · doi:10.1109/iscas.2015.7169109 · Cited by 11
“To account for different image resolution and viewing conditions, a multiscale SSIM (MS-SSIM) has been developed by combining the SSIM scores at different scales where the contrast and structure similarities are computed at each scale while the illumination similarity is computed only at the coarsest scale [10].”
MethodAutomatic parameter prediction for image denoising algorithms using perceptual quality features
Mittal, A., Moorthy, A. K., & Bovik, A. C. (2012) · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · doi:10.1117/12.912243 · Cited by 13
“We compare the quality of the resulting denoised image against the one obtained using the default implementation of the BM3D algorithm where quality of denoising is measured using MS-SSIM.”
MethodInfluentialSelf-citationJPEG2000 Image Compression on Solar EUV Images
Fischer, C. E., Müller, D., & Moortel, I. D. (2016) · Solar Physics · doi:10.1007/s11207-016-1038-3 · Cited by 11
“The structure term is determined by the correlation between two images u and v and measures the tendency of u and v to “vary together, [and is] thus an indication of structural similarity” (Wang, Simoncelli, and Bovik, 2003).”
BackgroundQuality assessment for image super-resolution based on energy change and texture variation
Fang, Y., Liu, J.-Y., Zhang, Y.-B., Lin, W., & Guo, Z.-M. (2016) · International Conference on Information Photonics · doi:10.1109/icip.2016.7532720 · Cited by 17
Objective quality assessment of retinal images based on texture features
Remeseiro, B., Mendonça, A. M., & Campilho, A. (2017) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn.2017.7966429 · Cited by 15
“The visual mean structural similarity (VMSSIM) index was proposed in [20], by adapting the well-known multi-scale structural similarity (MSSIM) index [8] to the visual system.”
BackgroundMethodColor-based structural similarity image quality assessment
Hassan, A. M., & Bashraheel, M. S. (2017) · doi:10.1109/icitech.2017.8079929 · Cited by 14
Kancharla, P., & Channappayya, S. S. (2018) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2018.8451296 · Cited by 15
“[7] demonstrated that autoencoders give superior results when trained using the MS-SSIM index [8] instead of the standard L1 or L2 distances.”
MethodInfluentialLi, S., Jin, W., Wang, X., Li, L., & Liu, M. (2018) · IEEE Access · doi:10.1109/access.2018.2873743 · Cited by 12
Miao, X., Chu, H., Liu, H., Yang, Y., & Li, X. (2019) · Signal Processing Image Communication · doi:10.1016/j.image.2019.08.013 · Cited by 12
Wang, X., Shao, F., Jiang, Q., Meng, X., & Ho, Y. (2020) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2020.2993942 · Cited by 12
Objective Image Quality Measures for Disparity Maps Evaluation
Merrouche, S., Andrić, M., Bondžulić, B. P., & Bujaković, D. M. (2020) · Electronics · doi:10.3390/electronics9101625 · Cited by 12
SR-ITM-GAN: Learning 4K UHD HDR With a Generative Adversarial Network
Zeng, H., Zhang, X., Yu, Z., et al. (2020) · IEEE Access · doi:10.1109/access.2020.3028584 · Cited by 11
“MS-SSIM [62] demonstrates multiscale structural similarity and supplies more flexibility than a single SSIM [61] value.”
BackgroundMethodPose Flow Learning From Person Images for Pose Guided Synthesis
Zheng, H., Chen, L., Xu, C., & Luo, J. (2020) · IEEE Transactions on Image Processing · doi:10.1109/tip.2020.3031108 · Cited by 25
“…evaluate the synthesis results, low-level metrics like Structural Similarity (SSIM) [43], Multi-scale Structural Similarity (MS-SSIM) [44] and perceptual-level metrics like Inception Score (IS) [34] and the Perceptual Image Patch Similarity Distance (LPIPS) [48] are measured on…”
BackgroundSemi-Supervised Automated Layer Identification of X-ray Tomography Imaged PCBs
Botero, U. J., Ganji, F., Asadizanjani, N., et al. (2020) · IEEE Physical Assurance and Inspection of Electronics (PAINE) · doi:10.1109/paine49178.2020.9337738 · Cited by 12
Improving Video Anomaly Detection Performance with Patch-level Loss and Segmentation Map
Yang, Y., Zhan, D., Yang, F. H., et al. (2020) · IEEE International Conference on Computer and Communications (ICCC) · doi:10.1109/iccc51575.2020.9345287 · Cited by 14
Attributes Aware Face Generation with Generative Adversarial Networks
Yuan, Z., Zhang, J., Shan, S., & Chen, X. (2021) · Proceedings - International Conference on Pattern Recognition/Proceedings/International Conference on Pattern Recognition · doi:10.1109/icpr48806.2021.9412022 · Cited by 16
“Finally we employ common metrics (e.g. BRISQUE, IS, FID and MS-SSIM) and an extra trained facial attribute predictor to do quantitative evaluations.”
MethodInfluentialVideo quality assessment with dense features and ranking pooling
Zhang, Y., He, L., Lu, W. B., Li, J., & Gao, X. (2021) · Neurocomputing · doi:10.1016/j.neucom.2021.06.026 · Cited by 12
Low Quality and Recognition of Image Content
Hofbauer, H., Autrusseau, F., & Uhl, A. (2021) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2021.3103394 · Cited by 14
Atta, M. E. E.-D., Ibrahim, D. K., & Gilany, M. I. (2021) · IEEE Sensors Journal · doi:10.1109/jsen.2021.3139560 · Cited by 12
Co-Net: A Collaborative Region-Contour-Driven Network for Fine-to-Finer Medical Image Segmentation
Liu, A., Huang, X., Li, T., & Ma, P. (2022) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv51458.2022.00177 · Cited by 17
“[33] proposed a new ms-ssim loss; it can not only concentrate on the similarity of the structure, but also provide different weights for different resolutions.”
BackgroundSkinner, J. G., Topping, G. J., Nagel, L., Heid, I., Hundshammer, C., Grashei, M., van Heijster, F. H. A., Braren, R., & Schilling, F. (2023) · Magnetic Resonance in Medicine · doi:10.1002/mrm.29676 · Cited by 15
Evaluation of video quality metrics on transmission distortions in H.264 coded video
Sedano, I., Kihl, M., Brunnstrom, K. E., & Aurelius, A. (2011) · IEEE International Symposium on Broadband Multimedia Systems and Broadcasting (BMSB) · doi:10.1109/bmsb.2011.5954960 · Cited by 12
A reduced-reference metric based on the interest points in color images
Nauge, M., Larabi, M., & Fernández-Maloigne, C. (2010) · Picture Coding Symposium · doi:10.1109/pcs.2010.5702578 · Cited by 12
“Nous conseillons les métriques PSNR, PNSR_HVSM [23], VIFP et MSSIM [24] pour garantir une certaine variété de résultats tout en maintenant une bonne corrélation avec le ha l-0 05 28 45 9, v er si on 1 - 21 O ct 2 01 0 jugement humain.”
BackgroundInfluentialOptical aberration correction for simple lenses via sparse representation
Cui, J., & Huang, W. (2017) · Optics Communications · doi:10.1016/j.optcom.2017.12.007 · Cited by 12
Neuromorphologicaly-preserving Volumetric data encoding using VQ-VAEPreprint
Tudosiu, P.-D. D., Varsavsky, T., Shaw, R., Graham, M., Nachev, P., Ourselin, S., Sudre, C., & Cardoso, M. (2020) · arXiv.org · doi:10.48550/arxiv.2002.05692 · Cited by 20
“, 2012) and Multi-Scale Structural Similarity (MS-SSIM) (Wang et al., 2003).”
MethodDeep learning for biomedical image reconstruction: a survey
Ben Yedder, H., Cardoen, B., & Hamarneh, G. (2021) · Artificial Intelligence Review · doi:10.1007/s10462-020-09861-2 · Cited by 107
“These metrics are the root mean squared error (RMSE) or normalized mean squared error, structural similarity index (SSIM) or its extended version multiscale structural similarity (Wang et al. 2003), and peak signal to noise ratio (PSNR).”
BackgroundUnsupervised learning inversion of seismic velocity models based on a multi‐scale strategy
Yang, S., Liu, B., Ren, Y., et al. (2025) · Geophysical Prospecting · doi:10.1111/1365-2478.13665 · Cited by 11
“For the MSSIM, the same scales and the weight of each scale are set, with reference to Wang et al. (2003).”
MSPFusion: A feature transformer for multidimensional spectral-polarization image fusion
Tong, G., Yao, X., Li, B., Fu, J., Wang, Y., Hao, J., Karim, S., & Yu, Y. (2025) · Expert Systems with Applications · doi:10.1016/j.eswa.2025.127079 · Cited by 15
ComAI: The Convergence of Communication and Artificial Intelligence
Zhang, P., Niu, K., Wang, X., Liu, Y., Liang, Z., Dong, C., Dai, J., Xu, X., Xu, W., Zhang, Z., et al. (2025) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2025.3608174 · Cited by 26
“…quality of reconstructed information (e.g., similarity, perception) such as bilingual evaluation understudy (BLEU) [107], multiscale structural similarity (MS-SSIM) [108] and learned perceptual image patch similarity (LPIPS) [109] under the same channel bandwidth ratios or signal-to-noise ratio.”
DM-FNet: Unified Multimodal Medical Image Fusion via Diffusion Process-Trained Encoder-Decoder
He, D., Li, W., Wang, G., Huang, Y., & Liu, S. (2025) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2025.3613156 · Cited by 21
“Finally, the final fused image I F is obtained via (12).”
IncepHoloRGB: multi-wavelength network model for full-color 3D computer-generated holography
Yu, X., Teng, Z., Fan, X., et al. (2025) · Opto-Electronic Advances · doi:10.29026/oea.2025.250130 · Cited by 11
Learning Dynamic 3D Geometry and Texture for Video Face Swapping
Otto, C., Naruniec, J., Helminger, L., Etterlin, T., Mignone, G., Chandran, P., Zoss, G., Schroers, C., Groß, M. H., Gotardo, P. F. U., et al. (2022) · Computer Graphics Forum · doi:10.1111/cgf.14705 · Cited by 12
Moghari, M. D., Sanaat, A., Young, N., et al. (2023) · Physics in Medicine and Biology · doi:10.1088/1361-6560/acdf3a · Cited by 11
Ambiguity in Solving Imaging Inverse Problems with Deep-Learning-Based Operators
Evangelista, D., Morotti, E., Piccolomini, E. L., & Nagy, J. (2023) · Journal of Imaging · doi:10.3390/jimaging9070133 · Cited by 10
“To evaluate and compare the deblurred images, we use visual inspection on a selected test image and exploit the Structural Similarity index (SSIM) [28] on the test set.”
MethodInfluentialTang, X., Zhao, J., Cui, G., et al. (2023) · Infrared Physics & Technology · doi:10.1016/j.infrared.2023.104796 · Cited by 11
Li, F., Zhou, Y., Chen, Y., Li, J., Dong, Z., & Tan, M. (2023) · Complex & Intelligent Systems · doi:10.1007/s40747-023-01185-2 · Cited by 12
An Improved GAN-Based Image Restoration Method for Imaging Logging Images
Cao, M., Feng, H., & Xiao, H. Y. (2023) · Applied Sciences · doi:10.3390/app13169249 · Cited by 15
Yu, D., He, Z., & Ma, L. (2023) · Journal of Building Engineering · doi:10.1016/j.jobe.2023.107600 · Cited by 11
Unsupervised SAR Despeckling by Combining Online Speckle Generation and Unpaired Training
Wang, C., Zheng, R., Zhu, J., He, X., & Li, X. (2023) · IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing · doi:10.1109/jstars.2023.3327180 · Cited by 12
“Clean optical images are easy to obtain and contain rich information about textures and details, so MS-SSIM [35] is introduced to make better use of these priors information of clean optical images, where the structural similarity (SSIM) is defined as where μ x , μ y , σ x , σ x , and σ xy are the…”
BackgroundZhao, Q., Ji, L., Su, Y., Zhao, Y., & Shi, J. (2023) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2023.3329115 · Cited by 12
“The multiscale SSIM (MS-SSIM) [50], which calculates the multiscale similarity of T and P , is adopted to compute the structure loss where MS-SSIM(.) represents the MS-SSIM function [50].”
MethodMu, B., Shao, F., Chen, H., et al. (2023) · Expert Systems with Applications · doi:10.1016/j.eswa.2023.122953 · Cited by 11
Pre- to post-contrast breast MRI synthesis for enhanced tumour segmentation
Osuala, R. O., Joshi, S., Tsirikoglou, A., Garrucho, L., Pinaya, W. H. L., Díaz, O., & Lekadir, K. (2024) · SPIE Medical Imaging: Image Processing · doi:10.1117/12.3006961 · Cited by 22
“The metric values alongside their standard deviation for (a) are PSNR ↑ =32.91(1.35), MS-SSIM ↑ =0.798(0.08), LPIPS ↓ =0.064(0.04), for (b) they are PSNR ↑ =32.42(1.68), MS-SSIM ↑ =0.780(0.07), LPIPS ↓ =0.084(0.05), and for (c) they resulted in PSNR ↑ =34.74(1.73), MS-SSIM ↑ =0.717(0.09), LPIPS ↓ =0.062(0.03).”
MethodInfluentialSrinath, S., Mitra, S., Rao, S., & Soundararajan, R. (2024) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv57701.2024.00010 · Cited by 32
Vats, A., Völgyes, D., Vermeer, M., Pedersen, M., Raja, K. B., Fantin, D. S. M., & Hay, J. A. (2024) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2024.3391391 · Cited by 14
“Furthermore, while the use of traditional metrics, such as structural similarity index (SSIM) [39] and multiscale SSIM (MS-SSIM) [40] are very common as loss functions for a reconstruction task such as this, we found that the negligible advantage obtained was negated by the computation overhead…”
MethodLight Flickering Guided Reflection Removal
Hong, Y., Chang, Y., Liang, J., Ма, Л., Huang, T., & Shi, B. (2024) · International Journal of Computer Vision · doi:10.1007/s11263-024-02073-z · Cited by 11
“M i x .”
Unsupervised MRI motion artifact disentanglement: introducing MAUDGAN
Safari, M., Yang, X., Chang, C., Qiu, R. L. J., Fatemi, A. B., & Archambault, L. (2024) · Physics in Medicine and Biology · doi:10.1088/1361-6560/ad4845 · Cited by 14
“Our proposed method was compared with two generative Pix2Pix and CycleGAN models to quantify remaining spatial artifact (NMSE and PSNR), spatial and texture similarity (SSIM, MS-SSIM, and MS-GMSD) between ground truth and motion-corrected images, and visual information similarity (VIF) between ground truth and motion-…”
InfluentialToward Efficient Video Compression Artifact Detection and Removal: A Benchmark Dataset
Lin, L., Wang, M., Yang, J., Zhang, K., & Zhao, T. (2024) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2024.3414549 · Cited by 11
“To demonstrate the availability of FPEA in VQA, several classic video quality assessment algorithms are employed including PSNR [53], SSIM [54], MS-SSIM [55] 3D-PSD [56], VIIDEO [57], TLVQM [29], VSFA [58] and GSTVQA [59].”
Nguyen, T. Y., Lam, P. D., Tinh, V. P., Le, D.-D., Nguyen, H. N., & Khoa, T. A. (2024) · IEEE Access · doi:10.1109/access.2024.3424569 · Cited by 13
“…indices: precision (specificity), recall (sensitivity), accuracy/range index (ACC), Dice coefficient (DC), area under the receiver operating characteristic curve, Jaccard loss (intersection over union [IoU]), focal loss, and multiscale structural similarity index measure (MS-SSIM) loss [49], [50].”
BackgroundSim, W., Yim, J.-S., & Lee, J. (2024) · Remote Sensing · doi:10.3390/rs16142623 · Cited by 16
“…of Model B was higher than that of Model A, both training and validation accuracies remained relatively stable compared to Model A. Similar to the results of previous studies, it was deemed effective to apply hybrid loss functions that combine three loss functions for imbalanced datasets [37–40].”
Su, H., Liu, L., Jeon, G., Wang, Z., Guo, T., & Gao, M. (2024) · IEEE Transactions on Consumer Electronics · doi:10.1109/tce.2024.3433432 · Cited by 11
“To enhance the contrast of the dehazing image in high-frequency regions, MS-SSIM loss [47] is introduced.”
Self-citationBS-Diff: Effective Bone Suppression Using Conditional Diffusion Models From Chest X-Ray Images
Chen, Z., Sun, Y., Ge, R., Qin, W., Pan, C., Deng, W., Liu, Z., Min, W., Elazab, A., Wan, X., et al. (2024) · IEEE International Symposium on Biomedical Imaging (ISBI) · doi:10.1109/isbi56570.2024.10635371 · Cited by 10
“The MS-SSIM [13] is derived from modeling the luminance, contrast, and structure of an image across multiple scales.”
MethodWTFusion: Wavelet-Assisted Transformer Network for Multisensor Image Fusion
Li, X., Li, Y., Chen, H., Sun, J., Wang, M., & Chen, L. (2024) · IEEE Sensors Journal · doi:10.1109/jsen.2024.3432143 · Cited by 11
Five Pitfalls When Assessing Synthetic Medical Images with Reference Metrics
Dohmen, M., Truong, T., Baltruschat, I. M., & Lenga, M. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-72744-3_15 · Cited by 20
“Its multi-scale variant MS-SSIM[10] calculates and combines multiple scores for different downscaled versions of the images.”
BackgroundGAI-Enhanced Robust Semantic Communication With Asymmetric Architecture
Ren, P., Wang, J., Hou, X., Chen, J., & Jiang, C. (2024) · IEEE Transactions on Cognitive Communications and Networking · doi:10.1109/tccn.2024.3520132 · Cited by 12
“For the reconstruction task, we use the peak signal-to-noise ratio (PSNR) and multi-scale structural similarity (MS-SSIM) as metrics [45].”
Explicit-NeRF-QA: A Quality Assessment Database for Explicit NeRF Model Compression
Xing, Y., Yang, Q., Yang, K., et al. (2024) · IEEE International Conference on Visual Communications and Image Processing (VCIP) · doi:10.1109/vcip63160.2024.10849926 · Cited by 12
“To evaluate the performance of the metrics, the Pearson linear correlation coefficient (PLCC) [36], the Spearman Rank Order Correlation Coefficient (SRCC) [37] and the Root Mean Square Error (RMSE) [36] are calculated.”
Yin, M., & Yang, J. (2025) · PLoS ONE · doi:10.1371/journal.pone.0314541 · Cited by 17
Denoising diffusion models with optimized quantum implicit neural networks for image generation
Zhang, J., Che, X., Fan, Y., Peng, S., Chen, G., Ma, Q., & Hu, J. (2025) · Future Generation Computer Systems · doi:10.1016/j.future.2025.107875 · Cited by 20
Physics-informed neural networks to reconstruct surface velocity field from drifter data
Bang, C., Altaher, A. S., Zhuang, H., et al. (2025) · Frontiers in Marine Science · doi:10.3389/fmars.2025.1547995 · Cited by 15
Subjective quality evaluation of compressed digital compound images
Yang, H., Fang, Y., Yuan, Y., & Lin, W. (2015) · Journal of Visual Communication and Image Representation · doi:10.1016/j.jvcir.2014.11.001 · Cited by 10
SSIM as global quality metric: A differential geometry view
Richter, T. (2011) · International Conference on Quality of Multimedia Experience (QoMEX) · doi:10.1109/qomex.2011.6065701 · Cited by 15
Motion trajectory based visual saliency for video quality assessment
Ma, L., Li, S., & Ngan, K. N. (2011) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2011.6116109 · Cited by 12
Exploring computation-communication tradeoffs in camera systems
Mazumdar, A., Moreau, T., Kim, S., Cowan, M., Alaghi, A., Ceze, L., Oskin, M., & Sathe, V. (2017) · IEEE International Symposium on Workload Characterization · doi:10.1109/iiswc.2017.8167775 · Cited by 14
“Here, we scaled bilateral grid sizes from 4 pixels-pergrid-vertex to 64 in each of three dimensions in a bilateral grid and evaluated the resulting impact on quality using MSSSIM [41].”
MethodQuality assessment of images undergoing multiple distortion stages
Athar, S., Rehman, A., & Wang, Z. (2017) · International Conference on Information Photonics · doi:10.1109/icip.2017.8296868 · Cited by 8
Self-citationScreen content image quality assessment based on the most preferred structure feature
Wu, J., Li, H., Xia, Z., & Xia, Z. (2018) · Journal of Electronic Imaging · doi:10.1117/1.jei.27.3.033025 · Cited by 11
Li, C., Zhang, Y., Wu, X., & Zheng, Y. (2018) · IEEE Access · doi:10.1109/access.2018.2877714 · Cited by 11
“Many IQA methods were developed, especially the full-reference (FR) IQA, which has the high correlation between subjective scores and objective quality prediction, such as VIF (Visual Information Fidelity) index, MAD (Most Apparent Distortion), and MS-SSIM (Multi-Scale Structural Similarity) [4], FSIM (Features Simila…”
MethodBlind DCT-based prediction of image denoising efficiency using neural networks
Rubel, O. S., Rubel, A. S., Lukin, V., & Egiazarian, K. (2018) · European Workshop on Visual Information Processing · doi:10.1109/euvip.2018.8611710 · Cited by 8
How Do Image Quality Metrics Perform on Contrast Enhanced Images?
Amirshahi, S., Kadyrova, A., & Pedersen, M. (2019) · European Workshop on Visual Information Processing · doi:10.1109/euvip47703.2019.8946143 · Cited by 8
“FR greyscale IQMs: Structural SIMilarity index (SSIM) [12], Multi-Scale SSIM (MSSSIM) [44], Feature SIMilarity for greyscale images (FSIM) [37], and Reducedreference Image Quality Metric for Contrast change (RIQMC) [45].”
MethodFull Reference Image Quality Assessment by Considering Intra-Block Structure and Inter-Block Texture
Yang, J.-F., Zhang, W., Li, X.-L., Zhou, T., & Ou, B. (2020) · IEEE Access · doi:10.1109/access.2020.3028282 · Cited by 12
“In a follow-up study, many novel methods extending SSIM are proposed [16]–[19], such as multi-scale SSIM, information weighted SSIM, gradient-based SSIM, and complex wavelet SSIM.”
MethodCast-Gan: Learning To Remove Colour Cast From Underwater Images
Li, C. Y., & Cavallaro, A. (2020) · Proceedings - International Conference on Image Processing · doi:10.1109/icip40778.2020.9191157 · Cited by 13
“If we denote the contrast comparison and the structure comparison as cm(·, ·) and sm(·, ·), respectively, the multi-scale structural similarity is defined as [33]”
BackgroundReconstructing the 3D digital core with a fully convolutional neural network
Li, Q., Chen, Z., He, J., Hao, S., Wang, R., Yang, H.-T., & Sun, H. (2020) · Applied Geophysics · doi:10.1007/s11770-020-0822-x · Cited by 12
“When training the FCN, we combine MS-SSIM and l 1 norm to form the loss function L Mix .”
MethodInfluentialPoint Cloud Distortion Quantification based on Potential Energy for Human and Machine Perception
Yang, Q., Chen, S., Xu, Y., et al. (2021) · eScholarship (California Digital Library) · Cited by 10
Contextual Convolutional Neural Networks
Duţă, I. C., Georgescu, M.-I., & Ionescu, R. T. (2021) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw54120.2021.00049 · Cited by 13
“As evaluation measures, we report the Inception Score (IS) [29] and the Fréchet Inception Distance (FID) [10] for CIFAR-10, and the multiscale structural similarity index measure (MS-SSIM) [33] and the Sliced Wasserstein Distance (SWD) [26] for CelebA.”
MethodNeural Reference Synthesis for Inter Frame Coding
Ding, D., Gao, X., Tang, C., & Ma, Z. (2021) · IEEE Transactions on Image Processing · doi:10.1109/tip.2021.3134465 · Cited by 14
ASSP: An adaptive sample statistics-based pooling for full-reference image quality assessment
Ling, Y., Zhou, F., Guo, K., & Xue, J. (2022) · Neurocomputing · doi:10.1016/j.neucom.2021.12.098 · Cited by 10
“The IQA models in the first group use handcrafted features for measuring the degradation, including the noise quality metric (NQM) [64], SSIM [10], MS-SSIM [11], FSIMc [15], IW-SSIM [12], VIF [25], IFC [24], GSM [14], GMSD [32], LLM [19] and SPSIM [21].”
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.