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 701–800 of 6,047
Objective Video Quality Assessment Combining Transfer Learning With CNN
Zhang, Y., Gao, X., He, L., Lu, W. B., & He, R. (2019) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2018.2890310 · Cited by 58
“In addition to the existing FR methods (peak signal to noise ratio, the visual information fidelity [3], SSIM [13], MS-SSIM [15], GMSD [36], and STMAD [5]), we also compare the GMSDFLOW that is the FR-VQA metric used in label generation, and the MS-SSIM combined with the optical flow (MSFLOW) in which the MS-SSIM inde…”
BackgroundMethodQuality Assessment for Video With Degradation Along Salient Trajectories
Wu, J., Liu, Y., Dong, W., Shi, G., & Lin, W. (2019) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2019.2908377 · Cited by 59
“To comprehensively illustrate the efficiency of the proposed FAST method, nine state-of-art FR VQA models (i.e., MOVIE [9], VQM_VFD [33], stRRED [34], Vis3 [11], FePVQ [35], FLOSIM [13], VMAF [36], SpEED [37] and Peng [14]), as well as two FR IQA models (GMSD [28] and MS-SSIM [38]) are compared with our model on five…”
MethodInfluentialHam, Y., & Kamari, M. (2019) · Automation in Construction · doi:10.1016/j.autcon.2019.102831 · Cited by 62
Learning To Blindly Assess Image Quality In The Laboratory And Wild
Zhang, W., Ma, K., Zhai, G., & Yang, X. (2020) · 2020 IEEE International Conference on Image Processing (ICIP) · doi:10.1109/icip40778.2020.9191278 · Cited by 55
“Our method performs significantly better than all competing models on all six databases, and is close to two full-reference models (MS-SSIM [26] and NLPD [27]).”
MethodSeamless Multimedia Delivery Within a Heterogeneous Wireless Networks Environment: Are We There Yet?
Trestian, R., Comşa, I.-S., & Tüysüz, M. F. (2018) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2018.2789722 · Cited by 61
“Table V lists the VQAMap rules for mapping the values of several objective metrics, such as PSNR, PSNRHuman Visual System (PSNR-HVS) [27] which considers the Contrast Sensitivity Function (CSF), PSNR-HVS-M [28] which along CSF it also considers the between-coefficient contrast masking of Discrete Cosine Transform basi…”
Background3D Panoramic Virtual Reality Video Quality Assessment Based on 3D Convolutional Neural Networks
Yang, J., Liu, T., Jiang, B., Song, H., & Lu, W. B. (2018) · IEEE Access · doi:10.1109/access.2018.2854922 · Cited by 54
Learning Raw Image Reconstruction-Aware Deep Image Compressors
Punnappurath, A., & Brown, M. S. (2020) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2019.2903062 · Cited by 54
“Therefore, following other recent works [11], [14], [16], [19], [24], we use MS-SSIM [20], which is a commonly employed perceptual fullreference imagemetric for comparing lossy image compression algorithms, for quantitative evaluation of our sRGB outputs.”
MethodInfluentialSemantic Structured Image Coding Framework for Multiple Intelligent Applications
Sun, S., He, T., & Chen, Z. (2020) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2020.3042517 · Cited by 67
Deep Homography for Efficient Stereo Image Compression
Deng, X., Yang, W., Yang, R., et al. (2021) · 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/cvpr46437.2021.00154 · Cited by 55
Studying osteoarthritis with artificial intelligence applied to magnetic resonance imaging
Calivá, F., Namiri, N. K., Dubreuil, M., Pedoia, V., Ozhinsky, E., & Majumdar, S. (2021) · Nature Reviews Rheumatology · doi:10.1038/s41584-021-00719-7 · Cited by 65
“Nature reviews | Rheumatology R e v i e w s 0123456789();: research projects such as FastMRI and challenges such as the Multi-channel MR Reconstruction (MC-MRRec) challenge 38,42,43 .”
BackgroundFang, J., Yang, J., Khader, A., & Xiao, L. (2024) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2024.3361553 · Cited by 90
“In accordance with the literature [64], [65], [66], [67], [68], [69], we use the mean PSNR (MPSNR), spectral Authorized licensed use limited to the terms of the applicable license agreement with IEEE.”
MethodImage quality evaluation of a new high-performance ring-gantry cone-beam computed tomography imager
Lustermans, D., Fonseca, G. P., Taasti, V. T., et al. (2024) · Physics in Medicine and Biology · doi:10.1088/1361-6560/ad3cb0 · Cited by 56
Sun, Y., Wang, L., Li, G., Lin, W., & Wang, L. (2024) · Nature Biomedical Engineering · doi:10.1038/s41551-024-01283-7 · Cited by 113
“As there are no ground truth reference images for the in vivo images, conventional reference-based metrics such as MSE, PSNR, SSIM, MS-SSIM, UQI and VIF cannot be used for quantitative evaluation.”
InfluentialNon-Shift Edge Based Ratio (NSER): An Image Quality Assessment Metric Based on Early Vision Features
Zhang, M., Mou, X., & Zhang, L. (2011) · IEEE Signal Processing Letters · doi:10.1109/lsp.2011.2127473 · Cited by 55
Perceptual Quality Assessment for Screen Content Images by Spatial Continuity
Fang, Y., Du, R., Zuo, Y., Wen, W., & Li, L. (2019) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2019.2951747 · Cited by 56
“To improve the performance of MSE and PSNR, many studies on visual quality assessment are proposed which model variable aspects in the characteristics of the HVS. Wang et al. proposed the well-known structural similarity (SSIM) index [4] and the multi-scale structural similarity (MS-SSIM) [5].”
MethodAVT-VQDB-UHD-1: A Large Scale Video Quality Database for UHD-1
Rao, R. R. R., Göring, S., Robitza, W., Feiten, B., & Raake, A. (2019) · IEEE International Symposium on Multimedia (ISM) · doi:10.1109/ism46123.2019.00012 · Cited by 74
AWNet: Attentive Wavelet Network for Image ISP
Dai, L., Liu, X., Li, C., & Chen, J. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-67070-2_11 · Cited by 76
“We also employ the structural similarity (SSIM) loss LSSIM [36] that is aiming to reconstruct the RGB images by enhancing on structural similarity index.”
Object-Centric Neural Scene RenderingPreprint
Guo, M., Fathi, A., Wu, J., & Funkhouser, T. A. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2012.08503 · Cited by 112
“For a quantitative evaluation, we report PSNR, SSIM [41], and LPIPS [44].”
MethodHigh resolution histopathology image generation and segmentation through adversarial training
Li, W., Li, J., Polson, J., Wang, Z., Speier, W. F., & Arnold, C. W. (2021) · Medical Image Analysis · doi:10.1016/j.media.2021.102251 · Cited by 61
Self-citationMultiscale transunet + + : dense hybrid U-Net with transformer for medical image segmentation
Wang, B., Wang, F., Dong, P., & Li, C. (2022) · Signal Image and Video Processing · doi:10.1007/s11760-021-02115-w · Cited by 58
“As the greater difference the regional distribution has, the higher the MS-SSIM value is [11], the MS-TransUNet + + will highly focuses on boundary regions.”
MethodInfluentialSi, J., Huang, B., Yang, H., et al. (2022) · IEEE Transactions on Image Processing · doi:10.1109/tip.2022.3164537 · Cited by 54
Zhao, L., Zhang, Y., & Cui, Y. (2022) · IEEE Sensors Journal · doi:10.1109/jsen.2022.3172132 · Cited by 54
Screen image quality assessment incorporating structural degradation measurement
Gu, K., Wang, S., Zhai, G., Ma, S., & Lin, W. (2015) · IEEE International Symposium on Circuits and Systems proceedings · doi:10.1109/iscas.2015.7168586 · Cited by 54
FISBLIM: A FIve-Step BLInd Metric for quality assessment of multiply distorted images
Gu, K., Zhai, G., Liu, M., Yang, X., Zhang, W., Sun, X., Chen, W., & Zuo, Y. (2013) · SiPS ... design and implementation - IEEE Workshop on Signal Processing Systems · doi:10.1109/sips.2013.6674512 · Cited by 60
Content-based image quality metric using similarity measure of moment vectors
Thung, K., Paramesran, R., & Lim, C.-L. (2011) · Pattern Recognition · doi:10.1016/j.patcog.2011.12.001 · Cited by 56
“Performance comparison of the proposed metric (S etr ) with other IQMs The performance of the proposed metric (S etr ) was compared with five existing objective quality metrics, namely the (1) PSNR [37], (2) SSIM [12,38], (3) MS-SSIM [13], (4) VIF [14,35], and (5) IQM model using Tchebichef moments by Wee et al. (Tch-…”
BackgroundInfluentialTian, S., Zhang, L., Morin, L., & Olivier Déforges, A. (2018) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2018.2875307 · Cited by 59
“in [50].”
Safety and Trustworthiness of Deep Neural Networks: A SurveyPreprint
Huang, X., Kroening, D., Kwiatkowska, M., Ruan, W., Sun, Y.-C., Thamo, E., Wu, M., & Yi, X. (2018) · arXiv.org · doi:10.48550/arxiv.1812.08342 · Cited by 53
Survey on QoE Assessment Approach for Network Service
Yang, M., Wang, S., Calheiros, R. N., & Yang, F. (2018) · IEEE Access · doi:10.1109/access.2018.2867253 · Cited by 54
“• MultiScale-SSIM (MS-SSIM) [49] is an extension of single-scale SSIM, originally proposed for still image and later extended to video.”
MethodInfluentialHuang, M., Shen, Q., Ma, Z., Bovik, A. C., Gupta, P., Zhou, R., & Cao, X. (2018) · IEEE Transactions on Image Processing · doi:10.1109/tip.2018.2865089 · Cited by 61
“measured MOS and PSNR, WS-PSNR [1], SSIM, MS-SSIM (Multi-Scale SSIM) [31] and VIFP (Visual Information Fidelity in the Pixel Domain) [32].”
Self-citationMultimodal MRI Reconstruction Assisted With Spatial Alignment Network
Xuan, K., Xiang, L., Huang, X., Zhang, L., Liao, S., Shen, D. G., & Wang, Q. (2022) · IEEE Transactions on Medical Imaging · doi:10.1109/tmi.2022.3164050 · Cited by 65
“In particular, we use the popular structural similarity (SSIM) [45], [46] as the reconstruction loss, i.”
MethodLocally Adaptive Structure and Texture Similarity for Image Quality Assessment
Ding, K., Liu, Y., Zou, X., Wang, S., & Ma, K. (2021) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3474085.3475419 · Cited by 81
“In this subsection, we test the gradient-based optimization performance of A-DISTS against four competing models - MAE, MS-SSIM [35], LPIPS [41], and DISTS [6] in the context of single image super-resolution.”
MethodNo-Reference Image Quality Assessment via Multibranch Convolutional Neural Networks
Pan, Z., Yuan, F., Wang, X., Xu, L., Shao, X. J., & Kwong, S. (2022) · IEEE Transactions on Artificial Intelligence · doi:10.1109/tai.2022.3146804 · Cited by 64
“There were many FR-IQA methods developed, such as SSIM [5], MS-SSIM [6], VIF [7], IFC [8], FSIM [9], and so on.”
MethodOptimizing Video Prediction via Video Frame Interpolation
Wu, Y., Wen, Q., & Chen, Q. (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/cvpr52688.2022.01729 · Cited by 75
“We use Multi-scale Structural Similarity Index Measure (MS-SSIM) [44] and LPIPS [48] as evaluation metrics.”
SiTH: Single-view Textured Human Reconstruction with Image-Conditioned Diffusion
Ho, H.-I., Song, J., & Hilliges, O. (2024) · 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/cvpr52733.2024.00058 · Cited by 113
“8.1.”
InfluentialPerformance evaluation of objective quality metrics for HDR image compression
Valenzise, G., Francesca De Simone, A., Lauga, P., & Frederic Dufaux, A. (2014) · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · doi:10.1117/12.2063032 · Cited by 55
Karam, L. J., & Zhu, T. (2015) · SPIE Proceedings · doi:10.1117/12.2080393 · Cited by 35
Blind quality assessment of videos using a model of natural scene statistics and motion coherency
Saad, M. A., & Bovik, A. C. (2012) · Asilomar Conference on Signals, Systems and Computers (ACSSC) · doi:10.1109/acssc.2012.6489018 · Cited by 64
“Since DMOS scores on VQA databases are usually only reported for complete video segments (10 seconds), we used the MS-SSIM index [6] applied on a frame basis against the reference video as a proxy for human scores.”
MethodSelf-citationNiu, Y., Zhang, H., Guo, W.-Z., & Ji, R. (2018) · IEEE transactions on circuits and systems for video technology (Print) · doi:10.1109/tcsvt.2016.2634590 · Cited by 46
An extensive performance evaluation of full-reference HDR image quality metrics
Zerman, E., Valenzise, G., & Frederic Dufaux, A. (2017) · Quality and User Experience · doi:10.1007/s41233-017-0007-4 · Cited by 58
“On the other hand, one can re-sort to metrics developed in the context of LDR imagery, such as simple arithmetic (PSNR, MSE), structural (SSIM [62] and its multiscale version [61]) and information-theoretic (e.g., VIF [55]) metrics.”
BackgroundMethodFull Reference Light Field Image Quality Evaluation Based on Angular-Spatial Characteristic
Meng, C., An, P., Huang, X., Yang, C., & Liu, D. (2020) · IEEE Signal Processing Letters · doi:10.1109/lsp.2020.2982060 · Cited by 56
Variable Bitrate Image Compression with Quality Scaling Factors
Chen, T., & Ma, Z. (2020) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp40776.2020.9053885 · Cited by 73
“Here D is the distortion measured by PSNR (Peak signal-to-noise ratio) or MS-SSIM [8] (Multiscale structural similarity), R is the bit rate, and λ is a hyperparameter which controls the tradeoff between R and D.”
BackgroundIntelligent Home 3D: Automatic 3D-House Design From Linguistic Descriptions Only
Chen, Q., Wu, Q., Tang, R., Wang, Y., Wang, S., & Tan, M. (2020) · 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/cvpr42600.2020.01264 · Cited by 70
“The generated interior textures are evaluated with popular image generation metrics such as Fr´echet Inception Distance (FID) [12] and Multi-scale Structural Similarity (MS-SSIM) [37].”
MethodInfluentialInterference Shaping for Improved Quality of Experience for Real-Time Video Streaming
Singh, S., Andrews, J. G., & Veciana, G. A. D. (2012) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2012.120811 · Cited by 57
“In [15], a full reference metric MS-SSIM was proposed that has been found to predict the perceptual quality with very high correlation.”
MethodInfluentialEffects of Compression on Remote Sensing Image Classification Based on Fractal Analysis
Chen, Z., Hu, Y., & Zhang, Y. (2019) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2019.2891679 · Cited by 54
“comparison include the full-reference IQA metrics PSNR, SSIM [49], MS-SSIM [42], VIF [50], and no-reference IQA metrics BLIINDS-II [43] and BRISQUE [36].”
BackgroundMethodDewarping Document Image by Displacement Flow Estimation with Fully Convolutional Network
Xie, G.-W., Yin, F., Zhang, X.-Y., et al. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-57058-3_10 · Cited by 52
“One of them is Multi-Scale Structural Similarity (MS-SSIM) [22] which evaluate the global similarity between the rectified document images and scanned images in multi-scale.”
BackgroundZhao, C., Wang, T., & Lei, B. (2020) · Neural Computing and Applications · doi:10.1007/s00521-020-05421-5 · Cited by 52
“[28] objective evaluation index can capture blur across multiple scales, comprehensively considering the influence”
BackgroundTask-Driven Semantic Coding via Reinforcement Learning
Li, X., Shi, J., & Chen, Z. (2021) · IEEE Transactions on Image Processing · doi:10.1109/tip.2021.3091909 · Cited by 56
“Therefore, many perceptual distortion metrics, such as structural similarity index (SSIM) [22] and multi scale structural similarity index (MS-SSIM) [23], have been proposed.”
BackgroundGAN-Based Facial Attribute Manipulation
Liu, Y., Li, Q., Deng, Q., Sun, Z., & Yang, M. (2023) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2023.3298868 · Cited by 72
“In addition, the (MS-SSIM) [132] is a commonly used variant of SSIM where the structural similarity is computed over multiple scales to incorporate image details at different resolutions.”
MethodDNeRV: Modeling Inherent Dynamics via Difference Neural Representation for Videos
Zhao, Q., Asif, M. S., & Ma, Z. (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/cvpr52729.2023.00202 · Cited by 95
“During training, we adopt Adam [17] as the optimizer with learning rate of 5×10−4 and cosine annealing learning rate schedule [47] and the batch size is set to 1.”
Depth Analogy: Data-Driven Approach for Single Image Depth Estimation Using Gradient Samples
Choi, S., Min, D., Ham, B., Kim, Y., Oh, C., & Sohn, K. (2015) · IEEE Transactions on Image Processing · doi:10.1109/tip.2015.2495261 · Cited by 53
“Third, instead of measuring residual errors in the depth domain, we measure the multi-scale structural similarity (MS-SSIM) [48] between the scaled version of test depth maps, αD∗ + β, and the ground truth ones.”
MethodInfluentialVideo Quality Assessment Based on Measuring Perceptual Noise From Spatial and Temporal Perspectives
Zhao, Y., Yu, L., Chen, Z., & Zhu, C. (2011) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2011.2157189 · Cited by 51
“MS-SSIM [65] 0.”
Pseudo-healthy synthesis with pathology disentanglement and adversarial learning
Xia, T., Chartsias, A., & Tsaftaris, S. A. (2020) · Medical Image Analysis · doi:10.1016/j.media.2020.101719 · Cited by 67
“This is achieved by evaluating their structural similarity to the input images outside the pathology regions, using a masked Multi-Scale Structural Similarity Index (MS-SSIM) 5 with window width of 11 (Wang et al., 2003).”
MethodHDR Environment Map Estimation for Real-Time Augmented Reality
Somanath, G., & Kurz, D. (2021) · 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/cvpr46437.2021.01114 · Cited by 79
“Hence retrieval of similar environment maps based on SSIM between rendered images is a baseline method to find environment maps that produce similar lighting on the objects.”
BackgroundInfluentialDSIC: Deep Stereo Image Compression
Liu, J. T., Wang, S., & Urtasun, R. (2019) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2019.00323 · Cited by 70
“Metrics: We report results on both peak signal-to-noise ratio (PSNR) and multi-scale structural similarity (MSSSIM) [46] as a function of bitrate.”
BackgroundMethodInfluentialWang, Y., Zhang, C., Lu, J., Bai, L., Zhao, Z., & Han, J. (2020) · IEEE Access · doi:10.1109/access.2020.2986130 · Cited by 54
Low-Light Homomorphic Filtering Network for integrating image enhancement and classification
Sobbahi, R. A., & Tekli, J. (2021) · Signal Processing Image Communication · doi:10.1016/j.image.2021.116527 · Cited by 59
TESR: Two-Stage Approach for Enhancement and Super-Resolution of Remote Sensing Images
Ali, A. M., Benjdira, B., Koubâa, A., Boulila, W., & El‐Shafai, W. (2023) · Remote Sensing · doi:10.3390/rs15092346 · Cited by 60
On the Role of ViT and CNN in Semantic Communications: Analysis and Prototype Validation
Yoo, H., Dai, L., Kim, S., & Chae, C. (2023) · IEEE Access · doi:10.1109/access.2023.3291405 · Cited by 63
Towards Scalable Neural Representation for Diverse Videos
He, B., Yang, X., Wang, H., Wu, Z., Chen, H., Huang, S., Ren, Y., Lim, S.-N., & Shrivastava, A. (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/cvpr52729.2023.00594 · Cited by 83
TransY-Net: Learning Fully Transformer Networks for Change Detection of Remote Sensing Images
Yan, T., Wan, Z., Zhang, P., Cheng, G., & Lu, H. (2023) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2023.3327253 · Cited by 57
“L SSIM loss considers a local neighborhood of each pixel [50].”
BackgroundSpectroformer: Multi-Domain Query Cascaded Transformer Network For Underwater Image Enhancement
Khan, M. R., Mishra, P., Mehta, N., Phutke, S. S., Vipparthi, S. K., Nandi, S., & Murala, S. (2024) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv57701.2024.00148 · Cited by 96
On the standardization of the JPEG XT image compression
Richter, T. (2013) · Picture Coding Symposium · doi:10.1109/pcs.2013.6737677 · Cited by 51
Image quality assessment based on DCT subband similarity
Balanov, A., Schwartz, A., Moshe, Y., & Peleg, N. (2015) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2015.7351172 · Cited by 63
“The MS-SSIM index considers the viewing-distance and accounts for the multi-scale 2105978-1-4799-8339-1/15/$31.00 ©2015 IEEE ICIP 2015 nature of the HVS.”
MethodInfluentialNew Verification Strategy for Finger-Vein Recognition System
Hsia, C. (2017) · IEEE Sensors Journal · doi:10.1109/jsen.2017.2772799 · Cited by 59
“[39] proposed a Multi-Scale SSIM (MS-SSIM) index, which has a higher compliance with human visual”
Duanmu, Z., Ma, K., & Wang, Z. (2018) · IEEE Transactions on Image Processing · doi:10.1109/tip.2018.2855403 · Cited by 58
Self-citationMC360IQA: The Multi-Channel CNN for Blind 360-Degree Image Quality Assessment
Sun, W., Luo, W., Min, X., Zhai, G., Yang, X., Gu, K., & Ma, S. (2019) · IEEE International Symposium on Circuits and Systems proceedings · doi:10.1109/iscas.2019.8702664 · Cited by 58
“We compare our model with 6 popular FR IQA models, which are PSNR, WS-PSNR [3], CPP-PSNR [4], SPSNR [2], SSIM [15], MS-SSIM [16].”
MethodLow-dose CT via deep CNN with skip connection and network-in-network
You, C., Yang, L., Zhang, Y., & Wang, G. (2019) · Developments in X-Ray Tomography · doi:10.1117/12.2534960 · Cited by 59
“e (HU) to the unit interval [0,1] before the images were fed to the network. Finally, we used three common image quality metrics: peak signal-to-noise ratio (PSNR), structural similarity index (SSIM) [18], and root-mean-square error (RMSE) to evaluate the denoised image quality. The visual inspection of our results in…”
MethodPredicting the Quality of Compressed Videos With Pre-Existing Distortions
Yu, X., Birkbeck, N., Wang, Y., Bampis, C. G., Adsumilli, B., & Bovik, A. C. (2021) · IEEE Transactions on Image Processing · doi:10.1109/tip.2021.3107213 · Cited by 60
“Early work on VQA models largely focused on referencebased models, beginning with the simple PSNR and advancing to more sophisticated FR frame-based models that better account for visual perception, including SSIM, MPQM [16], MS-SSIM [17], VIF, VSNR [18], FSIM [19], IW-SSIM [20], and DOG-SSIM [21], and spatial-tempora…”
BackgroundMethodSelf-citationDeep Implicit Volume Compression
Tang, D., Singh, S., Chou, P. A., Häne, C., Dou, M., Fanello, S., Taylor, J., Davidson, P., Guleryuz, O. G., Zhang, Y., et al. (2020) · 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/cvpr42600.2020.00137 · Cited by 54
“[67] and MS-SSIM [66]), our method preserves better spatio-temporal coherence and thus has better compression rate.”
MethodAV-Deepfake1M: A Large-Scale LLM-Driven Audio-Visual Deepfake Dataset
Cai, Z., Ghosh, S., Adatia, A. P., Hayat, M., Dhall, A., Gedeon, T., & Stefanov, K. (2024) · Proceedings of the 32nd ACM International Conference on Multimedia · doi:10.1145/3664647.3680795 · Cited by 59
3D GAN Inversion with Pose Optimization
Ko, J., Cho, K., Choi, D., Ryoo, K., & Kim, S. (2023) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv56688.2023.00298 · Cited by 70
“MSE, perceptual similarity metric using LPIPS [53] and structural similarity metric using MS-SSIM [49].”
MethodUtilizing Image Scales Towards Totally Training Free Blind Image Quality Assessment
Saha, A., & Wu, Q. M. J. (2015) · IEEE Transactions on Image Processing · doi:10.1109/tip.2015.2411436 · Cited by 54
“ηz is vector denoting the relative weights applied to the subbands based on the experiments of weights for different scales [34] such that ∑ z ηz = 1.”
MethodDiscrete wavelet transform-based structural similarity for image quality assessment
Yang, C., Wen-rui, G., & Po, L.-M. (2008) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2008.4711770 · Cited by 51
“In order to improve the assessment accuracy, SSIM has been expanded to multiscale SSIM [ 5 ] and CW-SSIM [6].”
MethodVQAMap: A Novel Mechanism for Mapping Objective Video Quality Metrics to Subjective MOS Scale
Moldovan, A.-N., Ghergulescu, I., & Muntean, C. H. (2016) · IEEE Transactions on Broadcasting · doi:10.1109/tbc.2016.2570002 · Cited by 58
“The MS-SSIM [32] metric is computed as the weighted average of SSIM for several downscaled versions of the video frame, in order to account for the variations of viewing conditions.”
BackgroundMethodInfluentialStatistical Nearest Neighbors for Image Denoising
Frosio, I., & Kautz, J. (2018) · IEEE Transactions on Image Processing · doi:10.1109/tip.2018.2869685 · Cited by 53
“We resort to a large set of image quality metrics to evaluate the filtered images: PSNR, SSIM [18], MSSSIM [19], GMSD [20], FSIM and FSIMC [21].”
MethodJoint-SRVDNet: Joint Super Resolution and Vehicle Detection Network
Mostofa, M., Ferdous, S. N., Riggan, B. S., & Nasrabadi, N. M. (2020) · IEEE Access · doi:10.1109/access.2020.2990870 · Cited by 53
“…results of our experiments using several objective image quality metrics such as Peak Signal-to-Noise ratio (PSNR), Multi-scale Structural Similarity (MSSIM) [57], Universal image Quality Index (UQI) [58] and Visual Information Fidelity (VIF) [59] on a validation subset of images for each dataset.”
MethodWickramaratne, S., & Mahmud, M. S. (2021) · Frontiers in Big Data · doi:10.3389/fdata.2021.659146 · Cited by 64
“Multi-scale structural similarity for image quality (MS-SSIM) is used to interpret the diversity of images, where a higher MS-SSIM score indicates a higher similarity between the two images (Sharma et al., 2018; Wang et al., 2003; Brooks et al., 2008).”
MethodYuan, F., Dai, N., Tian, S., Zhang, B., Sun, Y., Yu, Q., & Liu, H. (2020) · International Journal for Numerical Methods in Biomedical Engineering · doi:10.1002/cnm.3321 · Cited by 55
Towards Perceptually Optimized Adaptive Video Streaming-A Realistic Quality of Experience Database
Bampis, C. G., Li, Z., Katsavounidis, I. X., Huang, T.-Y., Ekanadham, C., & Bovik, A. C. (2021) · IEEE Transactions on Image Processing · doi:10.1109/tip.2021.3073294 · Cited by 62
“Our first experiment was to evaluate several well-known video quality and QoE metrics, including PSNR, PSNRhvs [71], SSIM [72], MS-SSIM [73], ST-MAD [74], ViS3 [24], VQM-VFD [75], V-BLIINDS [76], ST-RRED [77], VMAF [3] (version 0.6.1), SQI [10] and Video ATLAS [70].”
MethodInfluentialSelf-citationLiu, P., Zhang, C., Qi, H., Wang, G., & Zheng, H. (2022) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2022.3145815 · Cited by 50
“(3) We propose an MS-SSIM loss based on the SSIM (Structural Similarity) [15] evaluation index, which makes the model pay more attention to the image structure information.”
MethodInfluentialPhotoelastic Stress Field Recovery Using Deep Convolutional Neural Network
Tao, B., Wang, Y., Qian, X., Tong, X., He, F., Yao, W., Chen, B., & Chen, B. (2022) · Frontiers in Bioengineering and Biotechnology · doi:10.3389/fbioe.2022.818112 · Cited by 53
“In order to solve the problem of edge noise in the process of image generation, multiscale SSIM (MS-SSIM) is added to the loss function (Wang et al., 2003), which can effectively improve the impact of edge noise.”
MethodObjective Quality Assessment of Interpolated Natural Images
Yeganeh, H., Rostami, M. G., & Wang, Z. (2015) · IEEE Transactions on Image Processing · doi:10.1109/tip.2015.2456638 · Cited by 54
“This well fits into the category of reduced-reference (RR) IQA, where only partial information about the perfect-quality original image is accessible [18], [19].”
BackgroundSelf-citationZemliachenko, A., Kozhemiakin, R. A., Uss, M., Абрамов, С. К., Ponomarenko, N. N., Лукин, В. В., Vozel, B., & Chehdi, K. (2014) · Journal of Applied Remote Sensing · doi:10.1117/1.jrs.8.083571 · Cited by 54
Jung, Y. J., Kim, H. G., & Ro, Y. M. (2015) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2015.2430632 · Cited by 49
“In 2D QA models, the mean of SSIM index [26], multiscale SSIM (MS_SSIM) [44], and visual information fidelity (VIF) [45] were used.”
MethodPredicting the Quality of Images Compressed After Distortion in Two Steps
Yu, X., Bampis, C. G., Gupta, P., & Bovik, A. C. (2019) · IEEE Transactions on Image Processing · doi:10.1109/tip.2019.2922850 · Cited by 46
Self-citationPerceptual Deep Depth Super-Resolution
Voynov, O., Artemov, A., Egiazarian, V., Notchenko, A. V., Bobrovskikh, G., Burnaev, E. V., & Zorin, D. (2019) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2019.00575 · Cited by 51
“The examples include (1) the metrics based on simple vision models of structural similarity SSIM [52], FSIM [57], MSSIM [53], (2) based on a sophisticated model of low-level visual processing [35], or (3) on convolutional neural networks (see [58] for a detailed overview).”
MethodDeepSIC: Deep Semantic Image Compression
Luo, S., Yang, Y., Yin, Y., Shen, C., Zhao, Y., & Song, M. (2018) · Lecture notes in computer science · doi:10.1007/978-3-030-04167-0_9 · Cited by 62
“Metric To assess the visual quality of reconstructed images, we adopt Multi-Scale Structural Similarity Index Metric (MS-SSIM) [34] and Peak Signal-to-Noise Ratio (PSNR) [8] for comparing original, uncompressed images to compressed, degraded ones.”
MethodReal‐world super‐resolution of face‐images from surveillance cameras
Aakerberg, A., Nasrollahi, K., & Moeslund, T. B. (2021) · IET Image Processing · doi:10.1049/ipr2.12359 · Cited by 64
“Artificially Corrupted Images For our experiments on artificially corrupted images we evaluate the performance using three conventional IQA methods, PSNR, SSIM, and the later Multi Scale Structural Similarity index (MS-SSIM) [46].”
MethodRemodeling randomness prioritization to boost-up security of RGB image encryption
Al-Roithy, B. O., & Gutub, A. (2021) · Multimedia Tools and Applications · doi:10.1007/s11042-021-11051-3 · Cited by 52
Zhang, Y., Kang, L., Wong, I. H. M., Dai, W., Li, X., Chan, R., Hsin, M. K. Y., & Wong, T. T. W. (2021) · Advanced Science · doi:10.1002/advs.202102358 · Cited by 54
“In addition, structural similarity index measure (SSIM),[63] which predicts the perceived quality based on illuminance, contrast, and structure, is appended to the aforementioned loss functions.”
BackgroundDIODE: A Dense Indoor and Outdoor DEpth DatasetPreprint
Vasiljevic, I., Kolkin, N., Zhang, S., Luo, R., Wang, H., Dai, F. Z., Daniele, A. F., Mostajabi, M., Basart, S., Walter, M. R., et al. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1908.00463 · Cited by 429
“We use the same objective as in previous work [3], which consists of L1 loss, depth gradient, and structural similarity (SSIM) [33].”
MethodGCSBA-Net: Gabor-Based and Cascade Squeeze Bi-Attention Network for Gland Segmentation
Wen, Z., Feng, R., Liu, J., Li, Y., & Ying, S. (2020) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2020.3015844 · Cited by 50
Image synthesis: a review of methods, datasets, evaluation metrics, and future outlook
Baraheem, S. S., Le, T.-N., & Nguyen, T. (2023) · Artificial Intelligence Review · doi:10.1007/s10462-023-10434-2 · Cited by 63
Towards a Full-Reference Quality Assessment for Color Images Using Directional Statistics
Lee, D., & Plataniotis, K. N. (2015) · IEEE Transactions on Image Processing · doi:10.1109/tip.2015.2456419 · Cited by 54
Compressed image quality metric based on perceptually weighted distortion
Hu, S., Jin, L.-N., Wang, H.-L., Zhang, Y., Kwong, S., & Kuo, C.-C. J. (2015) · IEEE Transactions on Image Processing · doi:10.1109/tip.2015.2481319 · Cited by 43
“SSIM and MS-SSIM have similar performances on both databases, this is because both of them measure the structure distortion.”
BackgroundMethodInfluentialBernaś, T., Rajwa, B. P., Asem, E. K., & Robinson, J. P. (2005) · Journal of Biomedical Optics · doi:10.1117/1.2136313 · Cited by 54
Image Quality Assessment Based on Local Linear Information and Distortion-Specific Compensation
Wang, H.-L., Fu, J., Lin, W., Hu, S., Kuo, C.-C. J., & Zuo, L. (2017) · IEEE Transactions on Image Processing · doi:10.1109/tip.2016.2639451 · Cited by 27
“MS-SSIM [9], NQM [11], IW-SSIM [10], GSM [37],”
MethodDeep learning models for bone suppression in chest radiographs
Gusarev, M., Kuleev, R., Khan, A., Rivera, A. R., & Khattak, A. M. (2017) · IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology (CIBCB) · doi:10.1109/cibcb.2017.8058543 · Cited by 53
Full-Reference Image Quality Assessment by Combining Features in Spatial and Frequency Domains
Tang, Z., Zheng, Y., Gu, K., Liao, K.-Y., Wang, W., & Yu, M.-M. (2019) · IEEE transactions on broadcasting · doi:10.1109/tbc.2018.2871376 · Cited by 45
“Meanwhile, its extended versions, namely, multiscale SSIM (MSSIM) [14], measures structural information by embedding multiscale space in an image.”
BackgroundNo-Reference Quality Assessment for Screen Content Images Based on Hybrid Region Features Fusion
Zheng, L., Shen, L., Chen, J., An, P., & Luo, J. (2019) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2019.2894939 · Cited by 48
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.