A Multi-Scale Learning Local Phase and Amplitude Blind Image Quality Assessment for Multiply Distorted Images

Li, Chaofeng; Zhang, Yu; Wu, Xiaojun; Zheng, Yuhui

doi:10.1109/access.2018.2877714

Cited by 11 publications

(6 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As before, we provide the execution time of NR methods relative to the FR method PSNR for convenience in comparison with Table 17. Apart from the 14 NR methods being evaluated in this work, we have included the execution times of seven other well-known NR IQA methods in Table 27, which include: BLIINDS2 [166], DIIVINE [167], FRIQUEE [168], [169], Jet-LBP [170], MS-LQAF [171], NFERM [172], and TCLT [173]. We have not evaluated the performance of these methods because they take an excessive amount of time to estimate the quality of an image, and are infeasible for large-scale or real-time use.…”

Section: A Performance Of Nr Methodsmentioning

confidence: 99%

A Comprehensive Performance Evaluation of Image Quality Assessment Algorithms

Athar

Zhou

2019

IEEE Access

View full text Add to dashboard Cite

Image quality assessment (IQA) algorithms aim to predict perceived image quality by human observers. Over the last two decades, a large amount of work has been carried out in the field. New algorithms are being developed at a rapid rate in different areas of IQA, but are often tested and compared with limited existing models using out-of-date test data. There is a significant gap when it comes to large-scale performance evaluation studies that include a wide variety of test data and competing algorithms. In this work we aim to fill this gap by carrying out the largest performance evaluation study so far. We test the performance of 43 full-reference (FR), seven fused FR (22 versions), and 14 no-reference (NR) methods on nine subject-rated IQA datasets, of which five contain singly distorted images and four contain multiply distorted content. We use a variety of performance evaluation and statistical significance testing criteria. Our findings not only point to the top performing FR and NR IQA methods, but also highlight the performance gap between them. In addition, we have also conducted a comparative study on FR fusion methods, and an important discovery is that rank aggregation based FR fusion is able to outperform not only other FR fusion approaches but also the top performing FR methods. It may be used to annotate IQA datasets as a possible alternative to subjective ratings, especially in situations where it is not possible to obtain human opinions, such as in the case of large-scale datasets composed of thousands or even millions of images.

show abstract

Section: A Performance Of Nr Methodsmentioning

confidence: 99%

A Comprehensive Performance Evaluation of Image Quality Assessment Algorithms

Athar

Zhou

2019

IEEE Access

View full text Add to dashboard Cite

show abstract

“…The time for DNN based methods, EONSS, MEON [85], and WaDIQaM-NR [78], was evaluated both on the GPU and CPU, while that of all other methods was evaluated on the CPU only. It should be noted that the execution time of some other wellknown BIQA methods including BLIINDS2 [22], DIIVINE [21], FRIQUEE [28], MS-LQAF [127], NFERM [25], and TCLT [128], is even more than that of ILNIQE [17], making them infeasible for large-scale or real-time use, which is why we have not included them in our analysis. It can be seen from Table X that the execution time of EONSS is approximately 20 to 30 times faster than competitive BIQA methods, such as CORNIA [29], dipIQ [31], ILNIQE [17], and SISBLIM [14].…”

Section: B Performance Evaluation and Comparisonmentioning

confidence: 99%

Deep Neural Networks for Blind Image Quality Assessment: Addressing the Data Challenge

Athar¹,

Wang²,

Wang³

2021

Preprint

View full text Add to dashboard Cite

The enormous space and diversity of natural images is usually represented by a few small-scale human-rated image quality assessment (IQA) datasets. This casts great challenges to deep neural network (DNN) based blind IQA (BIQA), which requires large-scale training data that is representative of the natural image distribution. It is extremely difficult to create human-rated IQA datasets composed of millions of images due to constraints of subjective testing. While a number of efforts have focused on design innovations to enhance the performance of DNN based BIQA, attempts to address the scarcity of labeled IQA data remain surprisingly missing. To address this data challenge, we construct so far the largest IQA database, namely Waterloo Exploration-II, which contains 3,570 pristine reference and around 3.45 million singly and multiply distorted images. Since subjective testing for such a large dataset is nearly impossible, we develop a novel mechanism that synthetically assigns perceptual quality labels to the distorted images. We construct a DNN-based BIQA model called EONSS 1 , train it on Waterloo Exploration-II, and test it on nine subject-rated IQA datasets, without any retraining or fine-tuning. The results show that with a straightforward DNN architecture, EONSS is able to outperform the very state-of-the-art in BIQA, both in terms of quality prediction performance and execution speed. This study strongly supports the view that the quantity and quality of meaningfully annotated training data, rather than a sophisticated network architecture or training strategy, is the dominating factor that determines the performance of DNNbased BIQA models. (Note: Since this is an ongoing project, the final versions of Waterloo Exploration-II database, quality annotations, and EONSS, will be made publicly available in the future when it culminates.)

show abstract

“…The subjective consistency scores of the proposed method MD-IQA and 12 currently commonly used IQA methods for multiply distorted images in database LIVE-MD are shown in Table I. The 13 currently commonly used IQA methods include: PSNR (peak signal-to-noise ratio), WSNR (weighted signal-to-noise ratio), SSIM [19] (structural similarity), MS-SSIM [20] (multi-scale SSIM), CW-SSIM [21] (SSIM in complex wavelet domain), VIF [3], VIFp [3] (VIF in pixel domain), IFC [2], FSIM [22] (feature similarity), VSI [23] (visual saliency-induced index), as well as methods CM4 [7] CMSVR [8], and MS-LQAF [10], which combine multiple indexes. In the CM4, first, different exponential weights are assigned to the four indicators IFC, NQM, VSNR, and VIF, and then they are multiplied to obtain the image quality score.…”

Section: Comparison Of Subjective Consistency With Other Iqa Methods mentioning

confidence: 99%

“…Finally, the trained SVR model is used to obtain image quality scores. Chaofeng Li et al proposed [10] a multi-scale learning local phase and amplitude blind image quality assessment for multiply distorted images (MS-LQAF). First the distorted image is decomposed into three scales, and its phase congruency image (PCI), phase congruency covariance maximum image (PCCmax), phase congruency covariance minimum image (PCCmin) are constructed.…”

Section: Introductionmentioning

confidence: 99%

A Full-Reference Image Quality Assessment for Multiply Distorted Image based on Visual Mutual Information

Zhang¹,

Bai²,

Yan³

et al. 2019

jist

View full text Add to dashboard Cite

A full-reference image quality assessment (FR-IQA) method for multi-distortion based on visual mutual information (MD-IQA) is proposed to solve the problem that the existing FR-IQA methods are mostly applicable to single-distorted images, but the assessment result for multiply distorted images is not ideal. First, the reference image and the distorted image are preprocessed by steerable pyramid decomposition and contrast sensitivity function (CSF). Next, a Gaussian scale mixture (GSM) model and an image distorted model are respectively constructed for the reference images and the distorted images. Then, visual distorted models are constructed both for the reference images and the distorted images. Finally, the mutual information between the processed reference image and the distorted image is calculated to obtain the full-reference quality assessment index for multiply distorted images. The experimental results show that the proposed method has higher accuracy and better performance for multiply distorted images.

show abstract

A Multi-Scale Learning Local Phase and Amplitude Blind Image Quality Assessment for Multiply Distorted Images

Cited by 11 publications

References 28 publications

A Comprehensive Performance Evaluation of Image Quality Assessment Algorithms

A Comprehensive Performance Evaluation of Image Quality Assessment Algorithms

Deep Neural Networks for Blind Image Quality Assessment: Addressing the Data Challenge

A Full-Reference Image Quality Assessment for Multiply Distorted Image based on Visual Mutual Information

Contact Info

Product

Resources

About