Image Quality Assessment: From Error Measurement to Structural Similarity

By Zhou Wang, A. Bovik, Hamid R. Sheikh, Eero P. Simoncelli. 2004

1,226 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/0a895c20e6e115a36a26bf7d081eafcb10f97b09

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Image information and visual quality (2006, methodology): “: The VIF has one disadvantage when compared against PSNR or MSSIM: It has a higher computational complexity.”
Sparse Representation Based Image Interpolation With Nonlocal Autoregressive Modeling (2013, methodology): “Our experimental results on benchmark test images clearly demonstrate that the proposed NARM method outperforms much the classical bi-cubic interpolator [1-2], the representative edge-guided interpolators [3-5], and the recently developed SRM based image interpolation methods [6, 8] in term of PSNR,”
Comparative evaluation of despeckle filtering in ultrasound imaging of the carotid artery (2005, background): “The error summation in the form of the Minkowski metric, which is the norm of the dissimilarity between the original and the despeckled images [45]:”
An information fidelity criterion for image quality assessment using natural scene statistics (2005, methodology): “A number of limitations of HVS-based methods are discussed in [16].”
Image upsampling via imposed edge statistics (2007, methodology): “These differences are measured by a root mean square error (RMSE) after the images are normalized to have a zero mean and a unit variance and also by the Structural Similarity Image Quality (SSIQ) described in [Wang et al. 2004].”
RRED Indices: Reduced Reference Entropic Differencing for Image Quality Assessment (2012, methodology): “The mean-squared error (MSE) has been used as a quality metric for a very long time, owing to its simplicity, despite having a very poor correlation with human perception [1].”
Blind Image Quality Assessment: From Natural Scene Statistics to Perceptual Quality (2011, methodology): “Each BP subband is compared with the HP residual band (obtained from the steerable pyramid transform) using a windowed structural correlation [16].”
Gradient-Based Structural Similarity for Image Quality Assessment (2006, methodology): “As for GSSIM, each image is partitioned into overlapped 8h 8 blocks, for SSIM, the image is partitioned into overlapped 11h 11 blocks which is used in reference[ 6 ].”
Visual Importance Pooling for Image Quality Assessment (2009, methodology): “Plot of RMSE as a function of (cid:0) and (cid:2) for SS-SSIM with JPEG2000 distortion.”
A modified PSNR metric based on HVS for quality assessment of color images (2011, methodology): “Correlation of PSNRQ with different HVS based metrics like Edge Performance Index (EPI), Structural Correlation (SC), Mean Absolute Error (MAE) [14], CONTRAST (c(x,y)), and SSIM [5] has been calculated and it comes out to be 0.9953, 0.9971, 0.8267, 0.7385 and 0.9965 respectively.”
Video Primal Sketch: A Unified Middle-Level Representation for Video (2015, methodology): “For assessing the quality of the synthesized results quantitatively, we adopt two criteria for diﬀerent representations, rather than the traditional approach based on error-sensitivity as it has a number of limitations (Wang et al (2004)).”
Gradient Profile Prior and Its Applications in Image Super-Resolution and Enhancement (2011, methodology): “We also compute RMS, ERMS [5], and SSIM [47] to qualitatively measure the super-resolution results.”
Multi-scale dictionary for single image super-resolution (2012, methodology): “We also report the objective quality by the PSNR and SSIM indices[16] with the original HR images.”
Medical QoS provision based on reinforcement learning in ultrasound streaming over 3.5G wireless systems (2009, methodology): “Image quality (PSNR) > 36dB (QCIF - 144 × 176) > 35dB (CIF - 288 × 352) Image quality (SSIM [ 12 ]) > 0.9”
BM3D Image Denoising with Shape-Adaptive Principal Component Analysis (2009, methodology): “The PSNR and the mean structural similarity index map (MSSIM) [17] results of the proposed method are provided in Table I.”