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 501–600 of 6,047
Subjective and Objective Quality Assessment of Compressed Screen Content Images
Wang, S., Gu, K., Zhang, X., Lin, W., Zhang, L., Ma, S., & Gao, W. (2016) · IEEE Journal on Emerging and Selected Topics in Circuits and Systems · doi:10.1109/jetcas.2016.2598756 · Cited by 83
Screen Content Image Quality Assessment Using Multi-Scale Difference of Gaussian
Fu, Y., Zeng, H., Ma, L., Ni, Z.-K., Zhu, J., & Ma, K.-K. (2018) · IEEE transactions on circuits and systems for video technology (Print) · doi:10.1109/tcsvt.2018.2854176 · Cited by 63
Semi-Supervised Capsule cGAN for Speckle Noise Reduction in Retinal OCT Images
Wang, M., Zhu, W., Yu, K., Chen, Z., Shi, F., Zhou, Y., Ma, Y., Peng, Y., Bao, D., Feng, S., et al. (2021) · IEEE Transactions on Medical Imaging · doi:10.1109/tmi.2020.3048975 · Cited by 89
“In this article, referring to [58], M = 5, α1 = β1 = γ1 = 0.”
Confusing Image Quality Assessment: Toward Better Augmented Reality Experience
Duan, H., Min, X., Zhu, Y., Zhai, G., Yang, X., & Callet, P. L. (2022) · IEEE Transactions on Image Processing · doi:10.1109/tip.2022.3220404 · Cited by 99
Acceptability-Based QoE Models for Mobile Video
Song, W., & Tjondronegoro, D. (2014) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2014.2298217 · Cited by 88
“Many objective video quality assessment (VQA) metrics, such as structural similarity (SSIM) [1], multiscale SSIM (MS-SSIM) [2], and NTIA general model of video quality metric (VQM) [3], have been widely used as the QoE models.”
MethodInfluentialStill image/video frame lossy compression providing a desired visual quality
Zemliachenko, A., Lukin, V., Ponomarenko, N., Egiazarian, K., & Astola, J. (2015) · Multidimensional systems and signal processing · doi:10.1007/s11045-015-0333-8 · Cited by 78
HyperNeRF: A Higher-Dimensional Representation for Topologically Varying Neural Radiance FieldsPreprint
Park, K., Sinha, U., Hedman, P., Barron, J. T., Bouaziz, S., Goldman, D. B., Martin-Brualla, R., & Seitz, S. M. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2106.13228 · Cited by 77
“We measure visual quality with LPIPS [Zhang et al. 2018], MSSSIM [Wang et al. 2003], and PSNR. Note that because we are reconstructing dynamic scenes with a single moving camera, there can be ambiguities which cause slight differences from the true geometry or appearance of the scene.”
MethodAuto-Embedding Generative Adversarial Networks For High Resolution Image Synthesis
Guo, Y. F., Chen, Q., Chen, J., Wu, Q., Shi, Q., & Tan, M. (2019) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2019.2908352 · Cited by 86
“For quantitative evaluation, we use Frechet Inception Distance (FID) [36], Inception Score [7] and MS-SSIM [22], [47] to evaluate the generated samples.”
MethodLiu, P., Li, J., Wang, L., & He, G. (2022) · IEEE Geoscience and Remote Sensing Magazine · doi:10.1109/mgrs.2022.3165967 · Cited by 95
Joint Graph Attention and Asymmetric Convolutional Neural Network for Deep Image Compression
Tang, Z., Wang, H., Yi, X., Zhang, Y., Kwong, S., & Kuo, C. ‐. J. (2022) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2022.3199472 · Cited by 93
“For the comparison of rate-distortion performance, peak signal-to-noise ratio (PSNR) and MS-SSIM [16] are used as the quality metrics, and rate is denoted by bits per pixel (bpp).”
BackgroundMethodInfluentialCoNeRF: Controllable Neural Radiance Fields
Kania, K., Yi, K. M., Kowalski, M., Trzciniski, T., & Tagliasacchi, A. (2022) · 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/cvpr52688.2022.01807 · Cited by 112
“We measure Peak Signal-to-Noise Ratio (PSNR), Multi-scale Structural Similarity (MS-SSIM) [44], and Learned Perceptual Image Patch Similarity (LPIPS) [50] in Table 1.”
MethodStructural Similarity Based Image Quality Assessment
Wang, Z., Bovik, A. C., & Sheikh, H. R. (2017) · CRC Press eBooks · doi:10.1201/9781420027822-7 · Cited by 95
Self-citationImproved BM3D image denoising using SSIM-optimized Wiener filter
Hasan, M., & El-Sakka, M. R. (2018) · EURASIP Journal on Image and Video Processing · doi:10.1186/s13640-018-0264-z · Cited by 90
“There are other variants of recently proposed SSIM called multi-scale structural similarity (MS-SSIM) [19].”
BackgroundShao, F., Tian, W., Lin, W., Jiang, G., & Dai, Q. (2016) · IEEE Transactions on Image Processing · doi:10.1109/tip.2016.2538462 · Cited by 76
Efficient and Effective Context-Based Convolutional Entropy Modeling for Image Compression
Li, M., Ma, K., You, J., Zhang, D. D., & Zuo, W. (2020) · IEEE Transactions on Image Processing · doi:10.1109/tip.2020.2985225 · Cited by 104
“[29] exploited adaptive bit allocations and perceptual losses to boost the compression performance especially in terms of MS-SSIM [30].”
MethodInfluentialYang, J., Zhao, Y., Liu, J., Jiang, B., Meng, Q., Lu, W., & Gao, X. (2022) · IEEE Transactions on Cybernetics · doi:10.1109/tcyb.2020.3024627 · Cited by 78
“Classical FR-IQA metrics include the peak signal-to-noise ratio (PSNR), structural similarity index (SSIM) [11], multiscale SSIM (MS-SSIM) [12], feature similarity (FSIM) [13], visual information fidelity (VIF) [14], internal generative mechanism (IGM) [15], visual saliency-induced index (VSI) [16],…”
MethodMS-GraphSIM: Inferring Point Cloud Quality via Multiscale Graph Similarity
Zhang, Y., Yang, Q., & Xu, Y. (2021) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3474085.3475294 · Cited by 91
“One classic work is MS-SSIM, which generates the pyramid representation of images to emulate the perceptual characteristics of HVS. Inspired by MS-SSIM, we would like to present a generic multiscale representation of point clouds for PCQA.”
MethodInfluentialBereczky, M., Wieland, M., Krullikowski, C., Martinis, S., & Plank, S. (2022) · IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing · doi:10.1109/jstars.2022.3152127 · Cited by 99
“This is consistent with the results of Bai et al. , who reported good performance using a combination of the distribution based focal loss [50], the Structural SIMilarity loss [63], and the region-based IoU loss [64].”
ResultNo-Reference Video Quality Assessment Based on Artifact Measurement and Statistical Analysis
Zhu, K., Li, C., Asari, V., & Saupe, D. (2015) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2014.2363737 · Cited by 77
“The performances of PSNR, SSIM [32], MS-SSIM [33], and VIF [30] are also given in Table IV and Fig.”
BackgroundMethodInfluentialBlind Quality Assessment of Camera Images Based on Low-Level and High-Level Statistical Features
Liu, Y., Gu, K., Wang, S., Zhao, D., & Gao, W. (2018) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2018.2849602 · Cited by 77
“Following this philosophy, some SSIM variants methods were proposed successively and further improved the prediction accuracy of SSIM, e.g., information content weighting-SSIM (IWSSIM) [10], multi-scale SSIM (MS-SSIM) [11] and structural similarity weighting-SSIM (SW-SSIM) [12], etc.”
MethodTian, Y., Pang, G., Liu, F., Chen, Y., Shin, S. H., Verjans, J. W., Singh, R., & Carneiro, G. (2021) · Lecture Notes in Computer Science · doi:10.1007/978-3-030-87240-3_13 · Cited by 77
“Building upon the pre-trained encoder fθ(·) using the loss in (1), we fine-tune two state-of-the-art UAD methods, IGD [8] and F-anoGAN [37], and a baseline method, multi-scale structural similarity index measure (MS-SSIM)-based auto-encoder [48].”
BackgroundMethodInfluentialSaliency-induced reduced-reference quality index for natural scene and screen content images
Min, X.-K., Gu, K., Zhai, G.-T., Hu, M.-H., & Yang, X.-K. (2018) · Signal Processing · doi:10.1016/j.sigpro.2017.10.025 · Cited by 74
FSNet: An Identity-Aware Generative Model for Image-Based Face Swapping
Natsume, R., Yatagawa, T., & Morishima, S. (2019) · Lecture Notes in Computer Science · doi:10.1007/978-3-030-20876-9_8 · Cited by 74
“Subsequently, we calculate the absolute difference and MS-SSIM [32] between the input and swapped images.”
MethodOu, Y.-F., Xue, Y., & Wang, Y. (2014) · IEEE Transactions on Image Processing · doi:10.1109/tip.2014.2303636 · Cited by 77
Color Distribution Information for the Reduced-Reference Assessment of Perceived Image Quality
Redi, J., Gastaldo, P., Heynderickx, I. E. J., & Zunino, R. (2010) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2010.2087456 · Cited by 81
Stereoscopic image quality assessment method based on binocular combination saliency model
Liu, Y., Yang, J., Meng, Q., Lv, Z., Song, Z., & Gao, Z. (2016) · Signal Processing · doi:10.1016/j.sigpro.2016.01.019 · Cited by 75
Nehmé, Y., Florent Dupont, A., Farrugia, J., Callet, P. L., & Lavoué, G. (2020) · IEEE Transactions on Visualization and Computer Graphics · doi:10.1109/tvcg.2020.3036153 · Cited by 84
A survey on 360-degree video: Coding, quality of experience and streaming
Chiariotti, F. (2021) · Computer Communications · doi:10.1016/j.comcom.2021.06.029 · Cited by 97
“An interesting case is presented by the comparison between SSIM and MS-SSIM, whose relative performance is similar, but with a very high variance: MS-SSIM performs better on some datasets [137], but worse in others [111], and neither is clearly better in others [131].”
MethodInfluentialImage Quality Assessment for Magnetic Resonance Imaging
Kastryulin, S., Zakirov, J., Pezzotti, N., & Dylov, D. (2022) · IEEE Access · doi:10.1109/access.2023.3243466 · Cited by 26
“Our study includes the following 35 metrics: 17 FullReference IQMs (PSNR, SSIM [2], MS-SSIM [50], IWSSIM [51], VIF [35], GMSD [52], MS-GMSD [53], FSIM [36], VSI [54], MDSI [55], HaarPSI [56], Content and Style Perceptual Scores [57], LPIPS [58], DISTS [59], PieAPP [60], DSS [61]), 3 No-Reference IQMs (BRISQUE [4], PaQ…”
Towards the Perceptual Quality Evaluation of Compressed Light Field Images
Paudyal, P., Battisti, F., Sjöström, M., Olsson, R., & Carli, M. (2017) · IEEE Transactions on Broadcasting · doi:10.1109/tbc.2017.2704430 · Cited by 81
“However, there are no quality metrics specifically designed for light field images, and thus, in this work, we analyze the performances of the following 2D metrics: • Mean Squared Errors (MSE) [84] • Peak Signal-to-Noise Ratio (PSNR) [84] • Signal-to-Noise Ratio (SNR) [84] • Weighted SNR (WSNR) [85] • PSNR based on Huma…”
Extensive Benchmark and Survey of Modeling Methods for Scene Background Initialization
Jodoin, P., Maddalena, L., Petrosino, A., et al. (2017) · IEEE Transactions on Image Processing · doi:10.1109/tip.2017.2728181 · Cited by 76
“6) Multi-Scale Structural Similarity Index (MS-SSIM) [41], [43], [46]: This is the metric defined in [81], that uses structural distortion as an estimate of the perceived visual distortion of two images, evaluated at multiple scales.”
MethodStudy of 3D Virtual Reality Picture Quality
Chen, M., Jin, Y., Goodall, T., Yu, X., & Bovik, A. C. (2019) · IEEE Journal of Selected Topics in Signal Processing · doi:10.1109/jstsp.2019.2956408 · Cited by 83
Self-citationSelf-Supervised Pretraining for RGB-D Salient Object Detection
Zhao, X., Pang, Y., Zhang, L., Lu, H., & Ruan, X. (2022) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v36i3.20257 · Cited by 111
A Wavelet-Based Dual-Stream Network for Underwater Image Enhancement
Ma, Z., & Oh, C. (2022) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp43922.2022.9747781 · Cited by 116
“For the multi-color space fusion network, f S ( · ) , we use the Multi-scale Structural Similarity [22] (MS-SSIM) loss, L MS-SSIM , and (cid:96) 1 loss, L 1 , that can effectively capture the contrast and color errors, respectively: where G LL is the LL image obtained by applying DWT to the ground…”
MethodPerception-driven Accelerated Rendering
Weier, M., Stengel, M., Roth, T., Didyk, P., Eisemann, E., Eisemann, M., Grogorick, S., Hinkenjann, A., Kruijff, E., Magnor, M. A., et al. (2017) · Computer Graphics Forum · doi:10.1111/cgf.13150 · Cited by 84
“Several other perception-based image and video quality metrics have been developed, e.g., the Moving Picture Quality Metric (MPQM) [VdBLV96] and the Multi Scale Structural Similarity Index Metric (MS-SSIM) [WSB03] (based on SSIM) using structural similarities of the image at various scales.”
MethodToward Bridging the Simulated-to-Real Gap: Benchmarking Super-Resolution on Real Data
Köhler, T., Bätz, M., Naderi, F., Kaup, A., Maier, A., & Rieß, C. (2019) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2019.2917037 · Cited by 110
Live-cell imaging powered by computation
Shroff, H., Testa, I., Jug, F., & Manley, S. (2024) · Nature Reviews Molecular Cell Biology · doi:10.1038/s41580-024-00702-6 · Cited by 75
“In such difficult cases, high-resolution data can be computationally degraded to yield semi-synthetic low-resolution data, with high accuracy if the PSF and noise of the low-resolution data can be estimated 11,29,30 .”
Objective view synthesis quality assessment
Conze, P.-H., Robert, P., & Morin, L. (2012) · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · doi:10.1117/12.908762 · Cited by 78
“The functions , and denote respectively luminance, contrast and structure comparison functions and are defined in (2), (3) and (4).”
BackgroundVisCode: Embedding Information in Visualization Images using Encoder-Decoder Network
Zhang, P., Li, C., & Wang, C. (2020) · IEEE Transactions on Visualization and Computer Graphics · doi:10.1109/tvcg.2020.3030343 · Cited by 80
“where Lbce represents the BCEWithLogits loss [15] and Lssim denotes the structural similarity index (SSIM) loss [51].”
BackgroundDeep Learning-Based Inverse Scattering With Structural Similarity Loss Functions
Huang, Y., Song, R., Xu, K., Ye, X., Li, C., & Chen, X. (2020) · IEEE Sensors Journal · doi:10.1109/jsen.2020.3030321 · Cited by 75
“The structural similarity (SSIM) index [18] is a well-known metric to measure the perceptual distance of two images based on luminance, contrast and texture.”
BackgroundA Cross-Layer Design for Perceptual Optimization Of H.264/SVC with Unequal Error Protection
Khalek, A. A., Caramanis, C., & Heath, R. W. (2012) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2012.120802 · Cited by 86
“The MS-SSIM index is applied in conjunction with frame copy error concealment and temporal upsampling.”
MethodInfluentialSieveNet: A Unified Framework for Robust Image-Based Virtual Try-On
Jandial, S., Chopra, A., Ayush, K., Hemani, M., Kumar, A., & Krishnamurthy, B. (2020) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv45572.2020.9093458 · Cited by 73
Luo, W., Duan, S., & Zheng, J.-W. (2021) · IEEE Access · doi:10.1109/access.2021.3060947 · Cited by 79
Injected Infrared and Visible Image Fusion via $L_{1}$ Decomposition Model and Guided Filtering
Yan, H., Zhang, J., & Zhang, X. (2022) · IEEE Transactions on Computational Imaging · doi:10.1109/tci.2022.3151472 · Cited by 74
Structural Similarity Measure for Color Images
Hassan, A. M., & Bhagvati, C. (2012) · International Journal of Computer Applications · doi:10.5120/6169-8590 · Cited by 87
“In [7] the single-scale structural similarity index was extended to a multi-scale structural similarity index (MSSIM) that achieved better results than SSIM.”
BackgroundMethodPerceptually optimized image rendering
Laparra, V., Berardino, A., Ballé, J., & Simoncelli, E. P. (2017) · Journal of the Optical Society of America A · doi:10.1364/josaa.34.001511 · Cited by 81
“berg [9]. Our use of a simple physiologically-inspired model for assessing perceptual disortion oers opportunities for improvement (note that most image quality models are less physiologically based [30, 31, 19]). For example, the NLP can likely be improved by including relationships between channels, which could help…”
BackgroundMethodSelf-citationUnsupervised Image-to-Image Translation via Pre-Trained StyleGAN2 Network
Huang, J., Liao, J., & Kwong, S. (2021) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2021.3065230 · Cited by 87
“We also compared the proposed method with several stateof-the-art work quantitatively, where SSIM [48] was adopted as the matrix to evaluate reconstruction accuracy.”
Method3D-DCT based perceptual quality assessment of stereo video
Jin, L., Boev, A., Gotchev, A. P., & Egiazarian, K. (2011) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2011.6116175 · Cited by 72
Single-frame super resolution of remote-sensing images by convolutional neural networks
Tuna, Ç., Ünal, G., & Sertel, E. (2018) · International Journal of Remote Sensing · doi:10.1080/01431161.2018.1425561 · Cited by 77
“…quality criterion is based on the assumption that the human visual system is highly adapted for extracting structural information from the scene, and therefore a good approximation to perceived image quality can be provided by a measure of structural similarity (Wang, Simoncelli, and Bovik 2003).”
BackgroundGAN-based Virtual Re-Staining: A Promising Solution for Whole Slide Image AnalysisPreprint
Xu, Z., Huang, X., Moro, C. F., Bozóky, B., & Zhang, Q. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1901.04059 · Cited by 117
“SSIM has been used for assessing the image quality in many related studies [24] .”
MethodNon-local Attention Optimized Deep Image CompressionPreprint
Liu, H., Chen, T., Guo, P., Shen, Q., Cao, X., Wang, Y., & Ma, Z. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1904.09757 · Cited by 112
“MS-SSIM is reported to offer higher correlation with human perception, especially at low bit rates [15].”
BackgroundMethodMulti-Step Prediction of Occupancy Grid Maps With Recurrent Neural Networks
Mohajerin, N., & Rohani, M. (2019) · 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/cvpr.2019.01085 · Cited by 84
“To force the classifier inputs towards the target, a multi-scale Structural Similarity Index Metric [26] (SSIM) between the summation result and the target OGM is added to the cost.”
MethodExploiting user interactivity in quality assessment of point cloud imaging
Alexiou, E., & Ebrahimi, T. (2019) · International Conference on Quality of Multimedia Experience (QoMEX) · doi:10.1109/qomex.2019.8743277 · Cited by 71
“For geometry-plus-color metrics, the 3D models are projected on image planes and the PSNR, SSIM, MS-SSIM [18], and VIFP [19] algorithms are applied on the captured views.”
MethodInfluentialLiu, L., Zhang, H., Xu, X., Zhang, Z., & Yan, S. (2019) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2019.2944979 · Cited by 78
“We adopted multiscale structural similarity (MS-SSIM), a multiscale variant of the perceptual similarity metric that weights the relative importance between different scales [18], [41].”
MethodToward Practical Monocular Indoor Depth Estimation
Wu, C.-Y., Wang, J., Hall, M., Neumann, U., & Su, S. (2022) · 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/cvpr52688.2022.00379 · Cited by 115
Underwater Forward-Looking Sonar Images Target Detection via Speckle Reduction and Scene Prior
Long, H., Shen, L., Wang, Z., & Chen, J. (2023) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2023.3248605 · Cited by 94
Self-citationUSCFormer: Unified Transformer With Semantically Contrastive Learning for Image Dehazing
Wang, Y., Xiong, J., Yan, X., & Wei, M. (2023) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2023.3277709 · Cited by 80
Mohsen, S., Ali, A. M., El‐Rabaie, E. M., Elkaseer, A., Scholz, S. G., & Hassan, A. M. A. (2023) · IEEE Access · doi:10.1109/access.2023.3281529 · Cited by 89
UIQI: A Comprehensive Quality Evaluation Index for Underwater Images
Liu, Y., Gu, K., Cao, J., Wang, S., Zhai, G., Dong, J., & Kwong, S. (2023) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2023.3301226 · Cited by 84
“Later, some SSIM-type IQA measures, e.g., MS-SSIM [12], IW-SSIM [13], SC-SSIM [14], were developed from different concepts and further improved the prediction accuracy to some extent.”
MethodMulti-Realism Image Compression with a Conditional Generator
Agustsson, E., Minnen, D. C., Toderici, G. D., & Mentzer, F. (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/cvpr52729.2023.02138 · Cited by 142
“He et al . present the “ELIC” model [13], which is a state-of-the-art neural compression model for rate-distortion (MSE and MS-SSIM [44]) performance amongst practical methods.”
Review of objective video quality metrics and performance comparison using different databases
Vranješ, M., Rimac-Drlje, S., & Grgić, K. (2012) · Signal Processing Image Communication · doi:10.1016/j.image.2012.10.003 · Cited by 82
DIBR synthesized image quality assessment based on morphological pyramids
Sandic-Stankovic, D. D., Kukolj, D., & Callet, P. L. (2015) · 3DTV-Conference: The True Vision - Capture, Transmission and Display of 3D Video · doi:10.1109/3dtv.2015.7169368 · Cited by 71
A Perceptually Weighted Rank Correlation Indicator for Objective Image Quality Assessment
Wu, Q., Li, H., Meng, F., & Ngan, K. N. (2018) · IEEE Transactions on Image Processing · doi:10.1109/tip.2018.2799331 · Cited by 77
“For comparison, we use different rank correlation indicators to evaluate thirteen popular IQA algorithms, which include six full-reference (FR) metrics, i.e., PSNR [47], IWPSNR [48], SSIM [20], IWSSIM [12], MSSSIM [49], FSIM [50], and seven no-reference (NR) metrics, i.e., BIQI [51], BLIINDS II [52], BRISQUE [53], DII…”
BackgroundModeling the Time—Varying Subjective Quality of HTTP Video Streams With Rate Adaptations
Chen, C., Member, S., Choi, L. K., Member, S., Veciana, G. A. D., Caramanis, C., Heath, R. W., & Bovik, A. C. (2014) · Cited by 69
Self-citationVisual Importance and Distortion Guided Deep Image Quality Assessment Framework
Guan, J., Yi, S., Zeng, X., Cham, W.-K., & Wang, X. (2017) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2017.2703148 · Cited by 74
Learned perceptual image enhancement
Talebi, H., & Milanfar, P. (2018) · IEEE International Conference on Computational Photography (ICCP) · doi:10.1109/iccphot.2018.8368474 · Cited by 73
“Multi-scale SSIM [33] is an extension of SSIM which is computed at multiple image scales.”
MethodEfficient Variable Rate Image Compression With Multi-Scale Decomposition Network
Cai, C., Chen, L., Zhang, X., & Gao, Z. (2018) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2018.2880492 · Cited by 79
“In terms of MS-SSIM, our method can save 45.12% rate relative to JPEG, which outperforms all other codecs even including BPG. Compared with Toderici et al.’s and Theis et al.’s methods, our method has better performance measured by either MS-SSIM or PSNR.”
MethodInfluentialVideo Multimethod Assessment Fusion (VMAF) on 360VR Contents
Orduna, M., Díaz, C., Munoz, L., Perez, P., Benito, I., & Garcia, N. (2019) · IEEE Transactions on Consumer Electronics · doi:10.1109/tce.2019.2957987 · Cited by 94
Task-Aware Quantization Network for JPEG Image Compression
Choi, J., & Han, B. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-58565-5_19 · Cited by 94
“Percieved Quality We also trained the proposed quantization network using conventional distortion metrics such as pixel-wise MSE loss and MS-SSIM [40] loss.”
MethodDeep learning in medical image super resolution: a review
Yang, H., Wang, Z., Liu, X., Li, C., Xin, J., & Wang, Z. (2023) · Applied Intelligence · doi:10.1007/s10489-023-04566-9 · Cited by 90
“Objective function: The design of the objective function is highly related to the assessment metrics, such as L2 loss and MS-SSIM loss.”
BackgroundInfluentialSelf-citationTowards Generalized mmWave-based Human Pose Estimation through Signal Augmentation
Xue, H., Cao, Q., Miao, C., Ju, Y., Hu, H., Zhang, A., & Su, L. (2023) · ACM Annual International Conference on Mobile Computing and Networking (MobiCom) · doi:10.1145/3570361.3613302 · Cited by 95
“M denotes the MS-SSIM (Multi-scale Structural Similarity Index Measure) loss [43], which is calculated by applying the SSIM [42] loss on multiple scales of the input heatmaps and ranges from 0 (dissimilar) to 1 (the same).”
MethodFrequency compensated diffusion model for real-scene dehazing
Wang, J., Wu, S., Yuan, Z., Tong, Q., & Xu, K. (2024) · Neural Networks · doi:10.1016/j.neunet.2024.106281 · Cited by 86
Full Reference Image Quality Assessment Based on Saliency Map Analysis
Tong, Y., Konik, H., Cheikh, F. A., & Trémeau, A. (2010) · Journal of Imaging Science and Technology · doi:10.2352/j.imagingsci.technol.2010.54.3.030503 · Cited by 76
Computational Imaging for VLBI Image Reconstruction
Bouman, K. L., Johnson, M. D., Zoran, D., Fish, V. L., Doeleman, S. S., & Freeman, W. T. (2016) · 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/cvpr.2016.105 · Cited by 104
“nd PSNR, are sometimes uninformative in the context of highly degraded VLBI reconstructions. Therefore, we supplement the MSE metric with the perceptually motivated structural similarity (SSIM) index [42]. Since the absolute position of the emission is lost when using the bispectrum, we first align the reconstruction t…”
MethodYou Only Need Adversarial Supervision for Semantic Image SynthesisPreprint
Sushko, V., Schönfeld, E., Zhang, D., Gall, J., Schiele, B., & Khoreva, A. (2020) · International Conference on Learning Representations · doi:10.48550/arxiv.2012.04781 · Cited by 226
“To measure the variation in the multi-modal generation, we evaluate MS-SSIM (Wang et al., 2003) and LPIPS (Zhang et al., 2018b) between images generated from the same label map.”
MethodInfluentialPerceptual image quality assessment by independent feature detector
Chang, H.-W., Zhang, Q., Wu, Q., & Gan, Y. (2014) · Neurocomputing · doi:10.1016/j.neucom.2014.04.081 · Cited by 72
Single-image deblurring with neural networks: A comparative survey
Koh, J., Lee, J., & Yoon, S. (2020) · Computer Vision and Image Understanding · doi:10.1016/j.cviu.2020.103134 · Cited by 82
Learning to Remove Clutter in Real-World GPR Images Using Hybrid Data
Sun, H.-H., Cheng, W., & Fan, Z. (2022) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2022.3176029 · Cited by 87
“To train the CR-Net, a loss function combining MAE loss and MS-SSIM loss [52], [53] is used, which can be expressed as”
MethodA Review of Generative Models in Generating Synthetic Attack Data for Cybersecurity
Agrawal, G., Kaur, A., & Myneni, S. (2024) · Electronics · doi:10.3390/electronics13020322 · Cited by 102
No-Reference Image Sharpness Assessment Based on Maximum Gradient and Variability of Gradients
Zhan, Y., & Zhang, R. (2017) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2017.2780770 · Cited by 80
Thermal to Visible Facial Image Translation Using Generative Adversarial Networks
Wang, Z., Chen, Z., & Wu, F. (2018) · IEEE Signal Processing Letters · doi:10.1109/lsp.2018.2845692 · Cited by 68
“Generated visible images are also compared against ground truth visible images using structural similarity (SSIM) [21], multiscale (MS-SSIM) [22], peak signal to noise ratio (PSNR), and mean squared error (MSE).”
MethodSelf-citationLearning to synthesize: robust phase retrieval at low photon counts
Deng, M., Li, S., Goy, A., Kang, I., & Barbastathis, G. (2020) · Light Science & Applications · doi:10.1038/s41377-020-0267-2 · Cited by 81
“The NPCC has been shown [83] to be more effective in recovering fine features than conventional loss functions such as the Mean Square Error (MSE), Mean Absolute Error (MAE) and the Structural Similarity (SSIM) index [84, 85].”
MethodBridge-GAN: Interpretable Representation Learning for Text-to-Image Synthesis
Yuan, M., & Peng, Y. (2019) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2019.2953753 · Cited by 76
Gleason, C., Pavel, A., Liu, X., Carrington, P., Chilton, L. B., & Bigham, J. P. (2019) · Proceedings of the 21st International ACM SIGACCESS Conference on Computers and Accessibility · doi:10.1145/3308561.3353792 · Cited by 69
“To compute structural similarity, we use the Multi-Scale Structural Similarity (MS-SSIM) index [33] that considers the luminance, contrast, and structural similarity between image regions at various zoom levels.”
MethodImage Quality Assessment by Visual Gradient Similarity
Zhu, J., & Wang, N.-C. (2012) · IEEE Transactions on Image Processing · doi:10.1109/tip.2011.2169971 · Cited by 79
“The features already employed in FR IQAs are local variance and correlation coefficient [16], [17], visual information [19], phase [23], [24], singular value [20], and singular vector [21], or both [3], [22] in local singular value decomposition, Harris response [30], gradient [31]–[34], etc. Image…”
MethodZhou, W., Yu, L., Qiu, W., Zhou, Y., & Wu, M. (2017) · Information Sciences · doi:10.1016/j.ins.2017.02.049 · Cited by 69
“In this work, SVR with a radial basis function (RBF) kernel is adopted; this strategy employs the most widely used regression algorithm, and is highly effective for high-dimensional data pooling [22, 18, 19, 38, 10, 15, 34, 41].”
BackgroundMethodInfluentialUnified Binary Generative Adversarial Network for Image Retrieval and Compression
Song, J., He, T., Gao, L., Xu, X., Hanjalić, A., & Shen, H. T. (2020) · International Journal of Computer Vision · doi:10.1007/s11263-020-01305-2 · Cited by 73
“The experimental results obtained from the NUS-WIDE dataset are shown in Table 8, which demonstrates that our BGAN+ obtains the best performance in terms of MS-SSIM.”
MethodInfluentialVisual complexity analysis using deep intermediate-layer features
Saraee, E., Jalal, M., & Betke, M. (2020) · Computer Vision and Image Understanding · doi:10.1016/j.cviu.2020.102949 · Cited by 86
Conditional Entropy Coding for Efficient Video Compression
Liu, J. T., Wang, S., Ma, W.-C., Shah, M., Hu, R. M., Dhawan, P., & Urtasun, R. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-58520-4_27 · Cited by 74
“Moreover we plot the rate-distortion curve for multiscale structural similarity (MS-SSIM) [35], which is a commonly-used perceptual metric that captures the overall structural similarity in the reconstruction.”
MethodSimilarity and quality metrics for MR image-to-image translation
Dohmen, M., Klemens, M., Baltruschat, I. M., Truong, T., & Lenga, M. (2025) · Scientific Reports · doi:10.1038/s41598-025-87358-0 · Cited by 106
“For the LPIPS metric, an Alex-Net or VGG-architecture trained for classification on ImageNet was employed and linear weights were optimized by jointly training a second network to predict human judgements on a dataset comprising these judgements for distorted image patch pairs named Berkeley-Adobe…”
Zhang, Y., Liu, H., Yang, Y., Fan, X., Kwong, S., & Kuo, C. ‐. J. (2021) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2021.3076224 · Cited by 77
Learned Block-Based Hybrid Image Compression
Wu, Y., Li, X., Zhang, Z., Jin, X., & Chen, Z. (2021) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2021.3119660 · Cited by 86
Zhou, X., Fang, H., Fei, X., Shi, R., & Zhang, J. (2021) · IEEE Access · doi:10.1109/access.2021.3124814 · Cited by 70
“SSIM [59] (Structural Similarity) loss is initially designed in image quality assessment task, which can be used to acquire the structural information.”
MethodMPED: Quantifying Point Cloud Distortion Based on Multiscale Potential Energy Discrepancy
Yang, Q., Zhang, Y., Chen, S., Xu, Y., Sun, J., & Ma, Z. (2022) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2022.3213831 · Cited by 89
“Inspired by MSSIM [18], we extend PED to a multiscale form, i.e., MPED via MPEDC ¼ 1jCj X K2C PEDK; (9) in which C represents a collection of K, jCj represents the number of elements inC.”
BackgroundMethodResultInfluentialThe covariance environment defines cellular niches for spatial inference
Haviv, D., Remšík, J., Gatie, M. I., Snopkowski, C., Takizawa, M., Pereira, N., Bashkin, J., Jovanovich, S. B., Nawy, T., Chaligné, R., et al. (2024) · Nature Biotechnology · doi:10.1038/s41587-024-02193-4 · Cited by 104
New measurements reveal weaknesses of image quality metrics in evaluating graphics artifacts
Čadík, M., Herzog, R., Mantiuk, R., Myszkowski, K., & Seidel, H. (2012) · ACM Transactions on Graphics · doi:10.1145/2366145.2366166 · Cited by 73
“MS-SSIM is reported as the best-performer in many IQM comparison studies [Sheikh et al. 2006; Ponomarenko et al. 2009].”
MethodInfluentialBoundary-Aware Segmentation Network for Mobile and Web ApplicationsPreprint
Qin, X., Fan, D.-P., Huang, C., Diagne, C., Zhang, Z. V., Sant'Anna, A. C., Suarez, A. M., Jägersand, M., & Shao, L. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2101.04704 · Cited by 152
“The SSIM loss is a patch-level measure, which considers a local neighborhood of each pixel.”
BackgroundMethodInfluentialLearning to synthesise the ageing brain without longitudinal data
Xia, T., Chartsias, A., Wang, C., & Tsaftaris, S. A. (2021) · Medical Image Analysis · doi:10.1016/j.media.2021.102169 · Cited by 94
“We used standard definitions of mean squared error (MSE), peak signalto-noise ratio (PSNR) and structural similarity (SSIM) of window length of 11 (Wang et al., 2003) to evaluate the closeness of the predicted images to the ground-truth.”
Subjective and Objective Quality Assessment of High Frame Rate Videos
Madhusudana, P. C., Yu, X., Birkbeck, N., Wang, Y., Adsumilli, B., & Bovik, A. C. (2021) · IEEE Access · doi:10.1109/access.2021.3100462 · Cited by 65
Self-citationAn Introduction to Neural Data Compression
Yang, Y., Mandt, S., & Theis, L. (2023) · Foundations and Trends® in Computer Graphics and Vision · doi:10.1561/0600000107 · Cited by 162
“tics in order to avoid blocking artifacts [131].”
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.