A DCT Statistics-Based Blind Image Quality Index

By Michele A. Saad, A. Bovik, Christophe Charrier. IEEE Signal Processing Letters, 2010

413 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Media Technology

View Paper: https://www.semanticscholar.org/paper/2ee53ddcc131bd5070888729b3d39e81419c27c5

How researchers see this paper

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

In Computer Vision and Pattern Recognition

1. Dacheng Tao
2. Lei Zhang
3. A. Bovik
4. Ming-Hsuan Yang
5. Luc Van Gool
6. Kede Ma
7. Chen Change Loy
8. Radu Timofte

In Media Technology

1. A. Bovik
2. Kede Ma
3. Lei Zhang
4. Dacheng Tao
5. Zhou Wang
6. Jongyoo Kim
7. Sanghoon Lee
8. Guangtao Zhai

Citations in the corpus

• Blind Image Quality Assessment via Deep Learning (2015, methodology): “(c) BLIINDS-II.”
• Blind Image Quality Assessment Using Joint Statistics of Gradient Magnitude and Laplacian Features (2014, methodology): “trained a probabilistic model based on contrast and statistical features such as kurtosis and anisotropy in the DCT domain [9], [15].”
• UGC-VQA: Benchmarking Blind Video Quality Assessment for User Generated Content (2020, methodology): “Successful blind picture quality assessment (BIQA) models of this type have been developed in the wavelet (BIQI [48], DIIVINE [37], C-DIIVINE [49]), discrete cosine transform (BLIINDS [50], BLIINDS-II [51]), curvelet [52], and spatial intensity domains (NIQE [53], BRISQUE [38]), and have further bee”
• A no-reference metric for evaluating the quality of motion deblurring (2013, methodology): “We call the collection of each sharp image and its deblurring results a data group.”
• dipIQ: Blind Image Quality Assessment by Learning-to-Rank Discriminable Image Pairs (2017, methodology): “[7], [59] jointly modeled x and MOS using a multivariate Gaussian distribution and performed prediction by maximizing the conditional probability P(x|MOS).”
• Unsupervised feature learning framework for no-reference image quality assessment (2012, methodology): “Existing general-purpose NR-IQA algorithms can be broadly classified as (1) Natural scene statistics (NSS) based approaches [13, 14, 17, 18] and (2) Training-based approaches [7, 22, 29].”
• Objective quality assessment of multiply distorted images (2012, methodology): “The performance of existing objective IQA models such as [1, 2, 3, 4] has previously been measured on databases such as [5, 6] containing images corrupted by one of several pos­ sible distortions.”
• Blind Image Quality Assessment Using a General Regression Neural Network (2011, methodology): “In [ 42 ], another NR IQA algorithm called the BLIINDS index was proposed, which predicts image quality based on eight NSS features expressed as statistics of local DCT coefficients.”
• Learning without Human Scores for Blind Image Quality Assessment (2013, methodology): “As in many previous works [4, 14, 16], in our experiments we only consider 4 types of distortions that are common to the three databases: JPEG2000, JPEG, WN and GB.”
• Blind Image Quality Assessment: A Natural Scene Statistics Approach in the DCT Domain (2012, methodology): “In our prior work [18], we used such sample statistics (kurtosis, entropy, etc.,), without image modeling, to create a reasonably successful but preliminary blind IQA algorithm.”
• A Feature-Enriched Completely Blind Image Quality Evaluator (2015, methodology): “[13] proposed a BIQA model, called BLIINDS, by training a probabilistic model based on contrast and structure features extracted in the DCT domain.”
• RRED Indices: Reduced Reference Entropic Differencing for Image Quality Assessment (2012, background): “This approach can be refined by assuming that the distorted image is subjected to a set of possible distortions known a priori.”