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 4,001–4,100 of 6,047
Objective video quality assessment using the 3D dual-tree complex wavelet transform
Yonis, K. (2009) · Carleton University's Institutional Repository (MacOdrum Library, Carleton University) · doi:10.22215/etd/2009-09296 · Cited by 5
Weighted Chroma Downsampling and Luma-Referenced Chroma Upsampling for HDR/WCG Video Coding
Fu, Q., Jung, C., Yang, G., et al. (2019) · IEEE Access · doi:10.1109/access.2019.2911673 · Cited by 2
Cao, J., Li, Q., Li-ping, X., et al. (2022) · Journal of the Global Power and Propulsion Society · doi:10.33737/jgpps/151661 · Cited by 2
“Structural similarity loss measure, or SSIM (Wang et al., 2003), is a method to measure the similarity between two figures.”
MethodInfluentialQoS to QoE Mapping Based on Optimised BP Neural Networks
Wu, R., Li, S., Zhang, X., et al. (2023) · International Conference on Image, Vision and Computing (ICIVC) · doi:10.1109/icivc58118.2023.10270792 · Cited by 2
“QoE Prediction via QoS Parameters The Video Quality Algorithm (VQA) commonly used in the industry today are Multiscale Structural Similarity (MSSSIM) [7], Natural Image Quality Evaluator (NIQE) [8] and Video Quality Model (VQM) [9], etc.”
MethodLeveraging diffusion models for unsupervised out-of-distribution detection on image manifold
Liu, Z., Zhou, J., & Weinberger, K. Q. (2024) · Frontiers in Artificial Intelligence · doi:10.3389/frai.2024.1255566 · Cited by 3
A Image steganography Technique based on Denoising Diffusion Probabilistic Models
Xu, E., Cao, Y., Hu, L., & Wang, C. (2024) · Youth Academic Annual Conference of Chinese Association of Automation (YAC) · doi:10.1109/yac63405.2024.10598586 · Cited by 3
“We also quantify the evaluation metrics of our method and the StegGAN, which include the traditional evaluation metrics PSNR and SSIM[29], as well as the recently commonly used UQI[30], MS-SSIM[31] and VIF[32].”
MethodLiang, Z., Gabrieli, F., Pol, A., et al. (2024) · Remote Sensing · doi:10.3390/rs16173233 · Cited by 2
Kong, F., Huang, M., Tang, J., et al. (2024) · Journal of Applied Remote Sensing · doi:10.1117/1.jrs.18.036508 · Cited by 2
He, F.-L., Wang, Z., Yuan, S., Zhang, S., & Zhao, Z. (2024) · IEEE Access · doi:10.1109/access.2024.3465534 · Cited by 2
“L BCE ( k ) , L SSIM ( k ) and L IoU ( k ) denote the binary cross-entropy (BCE) loss [38], structural similarity index (SSIM) loss [39], and intersection-over-union (IoU) loss [40] applied to k .”
MethodSelf-citationON Annotation-Free Optimization of Video Coding for Machines
Windsheimer, M., Brand, F., & Kaup, A. (2024) · Proceedings - International Conference on Image Processing · doi:10.1109/icip51287.2024.10647318 · Cited by 2
AVAFN-adaptive variational autoencoder fusion network for multispectral image
Chu, W.-L., Tu, C.-C., & Jian, B. (2024) · Multimedia Tools and Applications · doi:10.1007/s11042-024-20340-6 · Cited by 2
“The metrics include Multiscale Structural Similarity (MSSIM) [25], Peak Signal To Noise Ratio (PSNR) [26], Mean-Square Error (MSE), Questions About Behavioral Function (QABF) [27], Naturalness Image Quality Evaluator (NIQE) [28], Quality with Contour Variation (QCV) [29], Average Gradient (AG)…”
Generalized Caputo–Fabrizio fractional operator: an application in image denoising
Khan, A. M., Gaur, S. B. C., & Suthar, D. L. (2024) · Applied Mathematics in Science and Engineering · doi:10.1080/27690911.2024.2434002 · Cited by 2
Ding, X., Song, H., & Zhang, X. (2024) · IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing · doi:10.1109/jstars.2024.3521415 · Cited by 4
“To ensure satisfactory visual effects, the calculation of multiscale structural similarity (MS-SSIM) [48] can be applied ) where L ( G ) represents the total loss of the entire network.”
Akhoondi, F., & Almasganj, F. (2025) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2024.107443 · Cited by 2
Image quality assessment by enabling inter-patch message passing via graph convolutional networks
Liu, Y., & Guo, J. (2025) · Neural Computing and Applications · doi:10.1007/s00521-024-10893-w · Cited by 5
Group-wise Semantic-enhanced Interaction Network for Remote Sensing Spatio-Temporal Fusion
Zhu, B., Hu, S., Song, H., et al. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10889185 · Cited by 3
“…as where the pixel loss L pixel and the structural loss L struct are defined as where I 2 f is the ground truth, ˆ I 2 f is the predicted image, MS-SSIM is the Multi-Scale Structural Similarity [28]. µ p and µ s are regularization parameters, which are set to 0.001 and 0.05, respectively.”
InfluentialDualNetIQ: Texture-Insensitive Image Quality Assessment with Dual Multi-Scale Feature Maps
Agamy, A., Mady, H., Esmaiel, H., Ayidh, A. A., Aly, A. M., & Abdel‐Nasser, M. (2025) · Electronics · doi:10.3390/electronics14061169 · Cited by 3
LPPSLF: a lightweight privacy-preserving split learning framework for smart surveillance systems
Wang, L., Chen, H., Zuo, L., & Liu, H. (2025) · Applied Intelligence · doi:10.1007/s10489-025-06489-z · Cited by 3
Secure and Robust Joint Source-Channel Coding With Semantic Clustering and Adversarial Purification
Huang, X., Zeng, L., Lu, Y., et al. (2025) · IEEE Transactions on Cognitive Communications and Networking · doi:10.1109/tccn.2025.3556777 · Cited by 2
“Commonly used distortion losses include mean squared error (MSE), peak signal-to-noise ratio (PSNR) [27], structural similarity index measure (SSIM) [28], and multi-scale SSIM (MS-SSIM) [29].”
Miao, Y., Hu, D., & Wang, Y. (2025) · IEEE Wireless Communications Letters · doi:10.1109/lwc.2025.3559583 · Cited by 2
“To evaluate performance, we adopt Multi-Scale Structural Similarity (MS-SSIM) [13] as a pixel-level quality metric and utilize Learned Perceptual Image Patch Similarity (LPIPS) as a semantic-level evaluation metric [14].”
Video Interpolation-Based Multi-Source Data Fusion Method for Laser Processing Melt Pool
Ren, H., Zhang, Y., Li, H., et al. (2025) · Applied Sciences · doi:10.3390/app15094850 · Cited by 2
DDVC: Deep Distributed Video Coding Using Quality Enhancement Network
Zhou, J., Ye, Z., Yi, X., Zhang, W., Lin, Q., & Xiang, J. (2025) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2025.3566706 · Cited by 3
“2) Evaluation Criteria: To measure the distortion of reconstructed frames, we use two evaluation metrics: PSNR and MS-SSIM [54].”
Semantic-Aided Parallel Image Transmission Compatible With Practical System
Xu, M., Wu, Y., Shi, Y., Xia, X., Debbah, M. A., Zhang, W., & Zhang, P. (2025) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2025.3566886 · Cited by 3
“Therefore, to conduct a more comprehensive assessment, we further include MS-SSIM [44] and LPIPS [45] as two perceptual quality metrics.”
Multi-Scale Semantic-Guidance Networks: Robust Blind Face Restoration Against Adversarial Attacks
Zhang, Z., Qi, X., Song, Z., Yang, Z. Q., Lu, J., Sun, M., Zhang, M., & Sun, Z. (2025) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2025.3570850 · Cited by 2
“Typically, in blind face restoration, the majority of BFR methods [50] utilize full-reference metrics, such as PSNR [51], SSIM [52], MS-SSIM [53], FID [9] and LPIPS, and no-reference metrics, such as BIQI [54], BRISQUE [55], SSEQ [56] and NIQE [57].”
FCA-Net: Accelerating stereo image compression through cascade alignment of side information
Xia, Y., Huang, Y., Chen, B., et al. (2025) · Pattern Recognition · doi:10.1016/j.patcog.2025.111799 · Cited by 2
Image Compression for Resource-Constrained AIoT System With Compressed Sensing
Chen, B., Huang, Y., Qiu, H., Xia, S., Wei, F., Wang, X., & Qiu, M. (2025) · IEEE Transactions on Systems Man and Cybernetics Systems · doi:10.1109/tsmc.2025.3571480 · Cited by 4
Chen, J., Chen, K., Zeng, H., Lin, Q., & Zhu, J. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2025.3577833 · Cited by 4
“The most used full reference objective quality assessment metrics are the PSNR, structural similarity (SSIM) [12], multiscale structural similarity (MS-SSIM) [19], visual information fidelity (VIF) based on information entropy [20], and information fidelity criterion (IFC) [21].”
Iterative Reconstruction with Dynamic ElasticNet Regularization for Nuclear Medicine Imaging
Kasai, R., & Otsuka, H. (2025) · Journal of Imaging · doi:10.3390/jimaging11070213 · Cited by 3
“[31].”
He, Q., Wang, X., Cai, C., Zhu, L., Zhang, B., Yang, G., Jian, W., Wang, L., & Zhen, X. (2025) · Complex & Intelligent Systems · doi:10.1007/s40747-025-02013-5 · Cited by 2
“…quality of the synthetic DBT syn and its similarity with the paired DBT real were quantified by four metrics, including the root mean square error (RMSE), the peak signal-to-noise ratio (PSNR), the universal image quality index (UQI) [80] and the multi-scale structural similarity (MS-SSIM) [81].”
Can turbidity improve underwater optical imaging in occlusion?
Huang, Y., Krishnan, G. S., & Javidi, B. (2025) · Optics Express · doi:10.1364/oe.566955 · Cited by 2
Micro-Image Domain View Synthesizer for Free Navigation With Focused Plenoptic Cameras
Fachada, S., Bonatto, D., Lafruit, G., et al. (2025) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2025.3590906 · Cited by 2
“Objective results (peak signal to noise ratio (PSNR), multi-scale structural similarity (MS-SSIM) [42] and learned percep-tualimagepatchsimilarity(LPIPS)[43])arereportedinTableIII withbestresultsinbold.”
Magnetogram filtering techniques for global coronal modelling
Murteira, J. M. L., Brchnelova, M., Lani, A., et al. (2025) · RAS Techniques and Instruments · doi:10.1093/rasti/rzaf030 · Cited by 2
Deep nested U-structure network with frequency attention for building semantic segmentation
Moghalles, K., Al-Huda, Z., al-Alimi, D., et al. (2025) · Scientific Reports · doi:10.1038/s41598-025-13890-8 · Cited by 2
“To evaluate errors, a standard binary cross entropy loss was employed; this approach utilizes a combination of SSIM, Dice, and crossentropy losses to capture thin object information at each scale effectively.”
InfluentialHigh-Quality SAR Ship Image Generation via Physically-Guided Diffusion Models
Chen, Z., Bi, X., Dai, M., et al. (2025) · International Symposium on Robotics, Artificial Intelligence and Information Engineering (RAIIE) · doi:10.1109/raiie65740.2025.11140550 · Cited by 2
“We introduce the MS-SSIM[7] index to assess the diversity of the 6, the comparison between real SAR ship images and those generated by ShipDM reveals a strikingly high degree of similarity.”
RT-GuIDE: Real-Time Gaussian Splatting for Information-Driven Exploration
Tao, Y., Ong, D., Murali, V., Spasojevic, I., Chaudhari, P., & Kumar, V. (2025) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2025.3615039 · Cited by 14
“We evaluate the Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index (SSIM) [32], Learned Perceptual Image Patch Similarity (LPIPS) [33] on the RGB images and Root Mean-Square-Error (RMSE) on the depth images.”
Tusher, A. S., Rahman, M. A., Islam, M. R., Bosak, S., & Hossain, M. J. (2025) · International Journal of Energy Research · doi:10.1155/er/8286945 · Cited by 3
“On the other hand, the multiscale structural similarity index measure (MS-SSIM) [51] is used as an example of the perceptual approach of objective metrics, which is widely accepted for its superior performance in resembling human perception.”
Asher, G. L., Wang, S., Zaki, B. I., Russo, G. A., Gill, G. S., Thomas, C. R., Prioleau, T., Li, Y., Zhang, R., Yan, Y., et al. (2025) · Frontiers in Oncology · doi:10.3389/fonc.2025.1672778 · Cited by 4
“Finally, C 1 , C 2 are constants applied as suggested by Wang et al (27).”
Multimodality Image Registration With Modality Distillation
Zhou, X., Zhao, J., Chen, L., Vivone, G., Liu, Y., Nie, J., & Liu, H. (2025) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2025.3621065 · Cited by 2
“To o ff er a more holistic evaluation of the discrepancies between the registered image and its original counterpart, the root-mean-square error (RMSE), structural similarity index (SSIM) [51], correlation coe ffi cient (CC), and visual information fidelity (VIF) [52] were selected as metrics.”
Qu, Q., Shen, Y., Chen, X., Chung, Y. Y., Cai, W., & Liu, T. (2025) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2025.3626550 · Cited by 8
“Our extensive evaluation encompassed 16 widely recognized full-reference quality assessment methodologies, including PSNR, SSIM [8], MS-SSIM [17], IW-SSIM [18], VIF [19], FSIM [20], GMSD [21], VSI [22], DSS [23], HaarPSI [24], MDSI [25], LPIPS [9], DISTS [26], STRRED [35], VMAF [36], and…”
Efficient Local-Global Collaboration Transcoding for JPEG AI
Wang, Y., Zhang, Z., Wu, Y., et al. (2025) · IEEE International Conference on Multimedia and Expo (ICME) · doi:10.1109/icme59968.2025.11208970 · Cited by 2
“…evaluate the transcoding performance based on the reconstruction quality of JPEG AI-encoded images, using the average Bjøntegaard-Delta rate (BD-Rate) [26] of several metrics defined by JPEG AI CTTC, including MS-SSIM [18], VIF [19], FSIM [20], NLPD [21], IW-SSIM [22], VMAF [23] and PSNR-HVS [24].”
Real-Scene Image Dehazing via Laplacian Pyramid-Based Conditional Diffusion Model
Wang, Y., Sun, J., Liu, H., Zhang, X.-P. S., & Wei, M. (2025) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2025.3632694 · Cited by 2
“During the training phase of LPCDM, in addition to the noise prediction loss L diff , we also incorporate L 1 loss and the SSIM [53] loss as additional constraints, specifically: Here, L low is combined with L diff for reconstructing the low-pass image, while L res is employed for aligning the…”
Wireless Video Semantic Communication With Decoupled Diffusion Multi-Frame Compensation
Xie, B., Wu, Y., Shi, Y., Feng, B., Zhang, W., Park, J., & Quek, T. Q. S. (2025) · IEEE Transactions on Communications · doi:10.1109/tcomm.2025.3634248 · Cited by 6
“4) Evaluation Metrics: We leverage the widely used pixel-wise metric peak signal-to-noise ratio (PSNR), perceptual-level multi-scale structural similarity (MS-SSIM) [44] and learned perceptual image patch similarity (LPIPS) [45] as measurements for the reconstructed image quality.”
Learning Deep ISP for High-Speed Cameras: Achieving DSLR-Quality Imaging Under High Frame Rates
Chen, H., Liu, T., Tu, T., Wei, W., Jin, Y., & Chen, E. (2025) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2025.3637771 · Cited by 2
“To assess image quality, for the paired test set, we adopt three widely used metrics: peak signal-to-noise ratio (PSNR), multi-scale structural similarity index (MS-SSIM) [78], and learned perceptual image patch similarity (LPIPS) [79].”
Model-Hopping Semantic Communication System for a Reliable and Secure Transmission
Jiang, H., Dong, C., Liang, H., Xu, X., Liu, Y., Zheng, Z., & Zhang, P. (2025) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2025.3616625 · Cited by 2
Efficient Seismic Data Acquisition Using 2-D Deep Compressed Sensing
Alhejaili, R., Alfarraj, M., Iqbal, N., Masood, M., & Al-Shaikhi, A. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2025.3632462 · Cited by 3
“Furthermore, instead of relying solely on the l 1 loss, the model incorporates multiscale structural similarity (MS-SSIM) [40] loss to preserve the structure of the data.”
HQWGAN: a hybrid quantum–classical Wasserstein generative adversarial network for image generation
Han, Q., Xu, Y., Wang, H., Gani, A., & Por, L. Y. (2025) · The Journal of Supercomputing · doi:10.1007/s11227-025-08172-z · Cited by 3
“We use the following four evaluation metrics to assess the performance of HQW-GAN: the Fréchet Inception Distance (FID) [36], Cosine Similarity [37], Peak Signal-to-Noise Ratio (PSNR) [38], and Structural Similarity Index Measure (SSIM) [39].”
InfluentialSPHARQ-Based Semantic Communication
An, W., Dong, C., Liang, H., et al. (2026) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2025.3649359 · Cited by 2
“3) Evaluation Metrics and Comparison Schemes: The image semantic recovery performance of the system is verified by using the PSNR [37], MS-SSIM [41] and IoU metrics, and the image semantic data transmission performance of the system is verified by using the average transmission number I of semantic…”
Noise-Tolerant Novel-View SAR Synthesis via Denoising Diffusion
Rahimi, A., & Yu, S. X. (2026) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2026.3654542 · Cited by 4
“Because speckle noise degrades the reliability of SSIM on SAR data, we follow prior works [66], [67] to first apply 7 × 7 Kuan filtering [68] [69] to evaluate perceptual quality across multiple spatial scales.”
A method for 3D reconstruction of ancient buildings driven by crowdsourced images
Liu, Y., Huo, L., Shen, W., & Shen, T. (2026) · npj Heritage Science · doi:10.1038/s40494-026-02346-5 · Cited by 3
“Although advanced metrics, including LPIPS 26 and MS-SSIM 27 , have made improvements, they usually require further adjustment to adapt to the multi-scale geometric features of ancient buildings.”
InfluentialDiffVC-OSD: One-Step Diffusion-based Perceptual Neural Video Compression Framework
Ma, W., & Chen, Z. (2025) · IEEE International Conference on Visual Communications and Image Processing (VCIP) · doi:10.1109/vcip67698.2025.11396819 · Cited by 15
Assessing image quality in photoacoustic imaging: A metric-based and deep learning-based evaluation
Brugge, M. V. D., Joseph, F. K., & Awasthi, N. (2026) · Photoacoustics · doi:10.1016/j.pacs.2026.100800 · Cited by 2
“MS-SSIM improves robustness against resolution and viewing condition changes [37].”
Evaluation of Objective Image Quality Metrics for High-Fidelity Image Compression
Mohammadi, S., Jenadeleh, M., Sneyers, J., Saupe, D., & Ascenso, J. (2026) · IEEE Access · doi:10.1109/access.2026.3669417 · Cited by 5
“• MS-SSIM [37]: An extension of SSIM that incorporates variations in image resolution and content by evaluating the structural similarity at multiple scales.”
GeoFWI: A Large Velocity Model Data Set for Benchmarking Full Waveform Inversion Using Deep Learning
Li, C., Shen, Y., Fomel, S. B., et al. (2026) · Journal of Geophysical Research Machine Learning and Computation · doi:10.1029/2025jh001037 · Cited by 2
Ultra-lightweight Neural Video Representation Compression
Kwan, H. M., Peng, T., Gao, G., et al. (2025) · Picture Coding Symposium · doi:10.1109/pcs65673.2025.11417542 · Cited by 2
“We measure the fidelity of reconstructed videos using both PSNR and MS-SSIM [39].”
DLF: Extreme Image Compression with Dual-Generative Latent Fusion
Xue, N., Jia, Z., Li, J., Li, B., Zhang, Y., & Lu, Y. (2025) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.01787 · Cited by 25
“Results for PSNR and MS-SSIM [43] are provided in our supplementary material.”
RQ-QGAN: recurrent quantum generative adversarial network with QLSTM-inspired hidden states
Qi, H., Xu, Y., Wang, H., Gani, A., & Por, L. Y. (2026) · The Journal of Supercomputing · doi:10.1007/s11227-026-08600-8 · Cited by 2
Corrected center-surround retinex: application to tone reproduction for high dynamic range images
Chen, Z.-S., & Tai, S. (2015) · International Journal of Computers and Applications · doi:10.1080/1206212x.2015.1076604 · Cited by 2
“There are some image quality assessment methods for HDR images such as the structural similarity (SSIM) approach,[28] multi-scale SSIM,[29,30] HDR visible difference predictor (HDR-VDP),[31] dynamic range-independent metric (DRIM),[32] and tone-mapped Table 1. the analysis result of the…”
MethodKim, H. G., Kang, J.-K., & Song, B. C. (2014) · IEIE Transactions on Smart Processing and Computing · doi:10.5573/ieiespc.2014.3.3.110 · Cited by 2
“For comparison in terms of the objective visual quality, an image quality assessment metric for an objective evaluation of the sharpness enhancement: multi-scale structural similarity (MSSSIM) can be employed [18].”
BackgroundVideo quality assessment based on motion dissimilarity
Vu, P. V., & Chandler, D. M. (2013) · Cited by 2
GMVP: gradient magnitude and variance pooling-based image quality assessment in sensor networks
Zhang, Z., & Liu, S. (2016) · EURASIP Journal on Wireless Communications and Networking · doi:10.1186/s13638-015-0477-0 · Cited by 1
Blind Image Quality Assessment Bases On Natural Scene Statistics And Deep Learning
Ge, D., & Song, J. (2016) · Advances in computer science research · doi:10.2991/iccsae-15.2016.174 · Cited by 3
Visual quality assessment of reconstructed background images
Shrotre, A., & Karam, L. J. (2016) · International Conference on Quality of Multimedia Experience (QoMEX) · doi:10.1109/qomex.2016.7498954 · Cited by 2
“Using IVQUEST, the following full reference objective metric algorithms were tested: Peak Signal to Noise Ratio (PSNR), structural similarity index (SSIM) [25], Multi Scale SSIM (MSSIM) [15], visual signal-to-noise ratio (VSNR) [26], visual information fidelity (VIF) [27], pixel-based VIF (VIFP) [27], universal qualit…”
MethodInfluentialEnhanced adaptive streaming over LTE-advanced wireless networks
Foerster, J. R., Oyman, Ö., Liao, Y., Rehan, M. M., & Taie, W. (2012) · Asilomar Conference on Signals, Systems and Computers (ACSSC) · doi:10.1109/acssc.2012.6489149 · Cited by 2
On the use of image quality estimators for improved JPEG2000 coding
Phan, T. D., Vu, P. V., & Chandler, D. M. (2012) · Asilomar Conference on Signals, Systems and Computers (ACSSC) · doi:10.1109/acssc.2012.6489017 · Cited by 2
Eliminating blocking artifacts in halftoning-based block truncation coding
Xu, Z., & Chan, Y. (2016) · European Signal Processing Conference (EUSIPCO) · doi:10.1109/eusipco.2016.7760381 · Cited by 2
Method to quality assessment of stereo images
Ma, J., & An, P. (2016) · Visual Communications and Image Processing · doi:10.1109/vcip.2016.7805563 · Cited by 2
Image Filter Based on Block Matching, Discrete Cosine Transform and Principal Component Analysis
Ramos, A. I. C., Riverón, E. M. F., Ramírez, P. M., et al. (2017) · Lecture notes in computer science · doi:10.1007/978-3-319-62434-1_34 · Cited by 2
“On the other hand there are some relatively novel image quality criterions, such as based on the properties of human visual system (PSNR-HVSM) [13, 14], feature similarity index (FSIM) [15], structure similarity index (SSIM) [16] or multiscale structure similarity index (MSSIM) [17].”
Fusion of Multimodal Color Medical Images Using Quaternion Principal Component Analysis
Nawaz, Q., Xiao, B., Li, W., et al. (2017) · Communications in computer and information science · doi:10.1007/978-981-10-6388-6_33 · Cited by 2
PA VIF: A passive aggressive visual information fidelity for full reference image quality assessment
Ma, X., & Jiang, X. (2017) · International Telecommunication Networks and Applications Conference (ITNAC) · doi:10.1109/atnac.2017.8215389 · Cited by 2
“, the structural similarity index (SSIM) [11], multi-scale SSIM (MS-SSIM) [12] and information weighted SSIM (IW-SSIM) [13].”
MethodFull Reference Objective Quality Assessment for Reconstructed Background Images
Shrotre, A., & Karam, L. J. (2018) · Journal of Imaging · doi:10.3390/jimaging4060082 · Cited by 3
“not been designed for assessing the quality of reconstructed background images. In [21] popular FR-IQA techniques including Peak Signal to Noise ratio (PSNR), Multi-scale Similarity metric (MS-SSIM) [19] and Color image Quality Measure (CQM) [20], were adopted for objectively comparing the performance of the different…”
BackgroundMethodInfluentialJang, J., Jang, H., Eo, T., Bang, K., & Hwang, D. (2018) · Applied Sciences · doi:10.3390/app8091688 · Cited by 2
A Study on Image Comparison Metrics for Atmospheric Scattering Phenomenon Rendering
Gałaj, T., & Wojciechowski, A. (2018) · Lecture notes in computer science · doi:10.1007/978-3-030-00692-1_4 · Cited by 2
Research on Subjective Quality Assessment of Light Field Images
Shan, L., Liu, D., An, P., et al. (2018) · doi:10.1145/3301326.3301364 · Cited by 3
Image Quality Assessment by Combining Masking Texture and Perceptual Color Difference Model
Tang, Z., Zheng, Y., Wang, W., & Liao, K. (2020) · KSII Transactions on Internet and Information Systems · doi:10.3837/tiis.2020.07.012 · Cited by 2
“Because of the great success of SSIM, its extended versions, called MS-SSIM [6] and IW-SSIM [7], have been introduced by multiscale and information content weighted SSIM.”
MethodA Subjective Study on Videos at Various Bit Depths
Mackin, A., Ma, D., Zhang, F., & Bull, D. (2021) · Picture Coding Symposium · doi:10.1109/pcs50896.2021.9477460 · Cited by 5
“The image quality metrics tested in this paper include: PSNR [37], SSIM [38], MS-SSIM [39], VSNR [40], VIF [41], and VMAF [42].”
MethodCopy–Move Forgery Detection by Using Key-Point-Based Harris Features and CLA Clustering
Rathi, K., & Singh, P. (2021) · Smart innovation, systems and technologies · doi:10.1007/978-981-33-4676-5_8 · Cited by 2
Understanding the perceived quality of video predictions
Somraj, N., Kashi, M. S., Arun, S. P., & Soundararajan, R. (2021) · Signal Processing Image Communication · doi:10.1016/j.image.2021.116626 · Cited by 3
DDocE: Deep Document Enhancement with Multi-scale Feature Aggregation and Pixel-Wise Adjustments
Bogdan, K. O. M., Megeto, G. A. S., Leal, R. W., Souza, G. D. M., Valente, A. C., & Kirsten, L. N. (2021) · Lecture notes in computer science · doi:10.1007/978-3-030-86198-8_17 · Cited by 2
Generative Adversarial Network for Multi Facial Attributes Translation
Liu, J., Liu, X.-Y., & Feng, Y. (2021) · IEEE Access · doi:10.1109/access.2021.3112895 · Cited by 2
“Multi-Scale Structural SIMilarity (MS-SSIM) [51] can measure the variety of generated images.”
BackgroundCross-Domain Feature Similarity Guided Blind Image Quality Assessment
Feng, C., Ye, L., & Zhang, Q. (2022) · Frontiers in Neuroscience · doi:10.3389/fnins.2021.767977 · Cited by 3
“…the computer program could access from the pristine (or reference) image, objective IQA could be categorized into full-reference IQA (FR-IQA) (Wang et al., 2003, 2004; Sheikh and Bovik, 2006; Larson and Chandler, 2010a; Li et al., 2011; Zhang et al., 2011, 2014; Liu et al., 2012; Chang et…”
BackgroundDeep Generative Modeling for Neural Compression
Tomczak, J. M. (2021) · Springer eBooks · doi:10.1007/978-3-030-93158-2_8 · Cited by 5
Spatial Quality Assessment of Pansharpened Images Based on Gray Level Co-Occurrence Matrix
Maleki, S. A., & Ghassemian, H. (2022) · Iranian Conference on Machine Vision and Image Processing · doi:10.1109/mvip53647.2022.9738763 · Cited by 2
Signal Reconstruction From Quantized Noisy Samples of the Discrete Fourier Transform
Goyal, M., & Kumar, A. (2022) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2022.3168127 · Cited by 2
“2Note that the output of the ESSIM and MS-SSIM depends on calibration parameters.”
MethodInfluentialTowards Patient Specific Reconstruction Using Perception-Aware CNN and Planning CT as Prior
Ghosh, S., Ernst, P., Rose, G., et al. (2022) · IEEE International Symposium on Biomedical Imaging (ISBI) · doi:10.1109/isbi52829.2022.9761462 · Cited by 3
Development of a Hairstyle Conversion System Based on Mask R-CNN
Jang, S.-G., Man, Q., & Cho, Y. I. (2022) · Electronics · doi:10.3390/electronics11121887 · Cited by 2
No-Reference Image Quality Assessment based on Quality Awareness Feature and Multi-task Training
Lai, L., Chu, J. H., & Leng, L. (2022) · Journal of Multimedia Information System · doi:10.33851/jmis.2022.9.2.75 · Cited by 3
Zhou, Y.-J., Liu, S.-Q., Xie, X., et al. (2022) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra46639.2022.9812075 · Cited by 2
Poster Abstract: Polar Code-based Approximate Communication System for Multimedia Web Pages
Shreshtha, A., Singla, P., & Sarangi, S. R. (2022) · ACM/IEEE International Conference on Information Processing in Sensor Networks (IPSN) · doi:10.1109/ipsn54338.2022.00069 · Cited by 2
Haptic Texture Rendering of 2D Image Based on Adaptive Fractional Differential Method
Hu, H., & Song, A. (2022) · Applied Sciences · doi:10.3390/app122312346 · Cited by 2
DBPNet: A dual-branch pyramid network for document super-resolution
Peng, J., Yi, Y., Yu, C., et al. (2022) · Pattern Recognition Letters · doi:10.1016/j.patrec.2022.12.013 · Cited by 2
Bilinear CNNs for Blind Quality Assessment of Fine-Grained Images
Liu, L., Yan, J., Fang, Y., & Jiang, W. (2022) · IEEE International Workshop on Multimedia Signal Processing (MMSP) · doi:10.1109/mmsp55362.2022.9948837 · Cited by 2
“We compare the proposed BFG model with twelve FR-IQA methods, including PSNR, PSNR-HVS [11], VSI [15], SSIM [12], MS-SSIM [13], IWS-SIM [14], FSIM [16], MAD [21], GSM [18], GMSD [17], IFC [19] and VIF [20] in [10], and six BIQA methods, including NIQE [23], ILNIQE [24], CNNIQA [25], WaDIQaM-NR…”
MethodM-3R: A Memory Based Approach for Streaming QoE Prediction under 3R settings
Ghosh, M., & Singhal, C. (2021) · IEEE International Conference on Advanced Networks and Telecommunications Systems · doi:10.1109/ants52808.2021.9936944 · Cited by 4
“We have experimentally analyzed how a combination of VQA metrics (involving FR, RR, and NR), [7] 0.7324 0.7281 0.5135 MS-SSIM (FR) [10] 0.7586 0.7343 0.4806 VMAF (FR) [13] 0.7724 0.7381 0.4534 GMSD (FR) [11] 0.6853 0.6649 0.5342 STRRED (RR) [9] 0.7753 0.7526 0.4524 NIQE (NR) [8] 0 [18] 0.7201…”
MethodInfluentialConvolutional Recurrent Neural Networks with Attention Mechanism for Streaming QoE Prediction
Zhang, X., Li, S., & Hu, F. (2022) · International Conference on Communication and Information Systems (ICCIS) · doi:10.1109/iccis56375.2022.9998164 · Cited by 3
“STSQ can be measured using off-the-shelf Video Quality Assessment (VQA) metrics, with the following options: peak signal-to-noise ratio PSNR, SSSIM[19], MS-SSIM[17], STRRED[16] and NIQE[18].”
Method3D22MX: Performance Subjective Evaluation of 3D/Stereoscopic Image Processing and Analysis
Escobar, J. J. M., del Villar, E. Y. A., Matamoros, O. M., & Chanona-Hernández, L. (2022) · Mathematics · doi:10.3390/math11010171 · Cited by 3
Kobayashi, D., Nakamura, K., Kitahara, M., et al. (2022) · IEICE Transactions on Information and Systems · doi:10.1587/transinf.2022edp7048 · Cited by 2
Learning-based Visual Compression
Ji, R., & Karam, L. J. (2023) · Foundations and Trends® in Computer Graphics and Vision · doi:10.1561/0600000101 · Cited by 3
Rahaman, J., Lukas, B., May, J. R., Puyana, C., Tsoukas, M. M., & Avanaki, K. (2023) · Photonics in Dermatology and Plastic Surgery · doi:10.1117/12.2651211 · Cited by 2
“However, the performance of MS-SSIM+L1 loss function outstrips MS-SSIM+L2 loss function based on the evaluation metrics, including PSNR, SSIM [57], MS-SSIM [58, 59], and MSE.”
MethodCombining Two Adversarial Attacks Against Person Re-Identification Systems
Andrade, E. D. O., Sampaio, I. G. B., Guerin, J. M., & Viterbo, J. (2023) · International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP) · doi:10.5220/0011623800003417 · Cited by 2
“The formulation of this loss function originates from an approach to the structural similarity image quality paradigm (Wang et al., 2003).”
Privacy Protection Against Reverse Image Search
Pratik, R., & Sendhil, R. (2023) · International Conference on Artificial Intelligence and Smart Energy (ICAIS) · doi:10.1109/icais56108.2023.10073803 · Cited by 2
“Finally, the intensity data into tanh space is translated as recommended in prior work [29] to guarantee that the input pixel intensity remains within the right range ([0, 255]).”
MethodImpact of Reproduction Size and Halftoning Method on Print Quality Perception
Pinćjer, I., Novaković, D., Nedeljković, U., et al. (2016) · Acta Polytechnica Hungarica · doi:10.12700/aph.13.3.2016.3.5 · Cited by 2
“The visibility threshold is set at four cycles per the degree of the visual angle [17].”
Background
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.