Image quality assessment via spatial‐transformed domains multi‐feature fusion

Yu, Miao; Yao, Zheng; Liao, Kaiyang; Tang, Zhisen

doi:10.1049/iet-ipr.2018.6417

Cited by 5 publications

(6 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Similar to refs. [1–39], we use two widely used metrics to measure the prediction accuracy of IQA models. One is the Pearson linear correlation coefficient (PLCC).…”

Section: Methodsmentioning

confidence: 99%

“…Hence, it is urgent to develop an effective computer algorithm to automatically predict image quality scores, which is known as the objective IQA method. According to the use of reference information, the objective IQA can be divided into three categories: full‐reference IQA (FR‐IQA) [1–5], reduce‐rReference IQA (RR‐IQA) [6], and no‐reference IQA (NR‐IQA) [7–10]. With full or partial reference information for comparison, these FR‐IQA and RR‐IQA methods have achieved great prediction accuracy over the past decades.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

CPDINet: Blind image quality assessment via a content perception and distortion inference network

Shao

Liu

Zhang

2022

IET Image Processing

View full text Add to dashboard Cite

Nowadays, it is still challenging for the blind image quality assessment (BIQA) to accurately predict quality scores of distorted images, since distorted images have rich content information and complex distortions. To solve these problems, a content perception and distortion inference network for BIQA is proposed, which divides IQA task into content perception and distortion inference processes. Since humans try to understand image content before perceiving quality scores, a content feature extractor is designed to explore content information in an image to deal with the content variation problem. To handle the distortion diversity problem, a distortion feature extractor is proposed to capture distortion features in images. Because extracted content features and distortion ones have different characteristics, attention-based fusion blocks to fuse multi-scale content features and distortion ones as guidance to selectively enhance important features based on calculated weight scores are proposed. With fused features, a quality prediction module is designed to regress multi-scale features to quality scores. Experiments are performed on six public IQA datasets, including LIVE, CSIQ, TID2013, LIVEC, KonIQ-10k, and SPAQ. Experimental results show that the proposed method can effectively predict quality scores for both synthetically and authentically distorted images than its peers, including the state-of-the-art methods.

show abstract

“…Similar to refs. [1–39], we use two widely used metrics to measure the prediction accuracy of IQA models. One is the Pearson linear correlation coefficient (PLCC).…”

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

CPDINet: Blind image quality assessment via a content perception and distortion inference network

Shao

Liu

Zhang

2022

IET Image Processing

View full text Add to dashboard Cite

show abstract

“…Image enhancement algorithms, aiming to increase useful image information and reduce invalid information such as burr noise through pixel-level processing of images, eventually provide more effective image features for the human eye to acquire information or for subsequent computer processing. In the evolution of image enhancement algorithms, the anisotropic diffusion model image enhancement algorithm based on partial differential equations inherited and optimized the local processing enhancement algorithm, where the traditional Perona-Malik model can improve image contrast, increase image details, and reduce noise by combining with gradient calculation [5]. However, since this method smoothes the detail part in the image enhancement process, resulting in more loss of detail information in the image during enhancement, there is a need to add the processing of retaining the detail part for this algorithm.…”

Section: Introductionmentioning

confidence: 99%

Multifeature Contrast Enhancement Algorithm for Digital Media Images Based on the Diffusion Equation

Wang

Yuan

2022

Advances in Mathematical Physics

View full text Add to dashboard Cite

This paper studies the processing of digital media images using a diffusion equation to increase the contrast of the image by stretching or extending the distribution of luminance data of the image to obtain clearer information of digital media images. In this paper, the image enhancement algorithm of nonlinear diffusion filtering is used to add a velocity term to the diffusion function using a coupled denoising model, which makes the diffusion of the original model smooth, and the interferogram is solved numerically with the help of numerical simulation to verify the denoising processing effect before and after the model correction. To meet the real-time applications in the field of video surveillance, this paper focuses on the optimization of the algorithm program, including software pipeline optimization, operation unit balancing, single instruction multiple data optimization, arithmetic operation optimization, and onchip storage optimization. These optimizations enable the nonlinear diffusion filter-based image enhancement algorithm to achieve high processing efficiency on the C674xDSP, with a processing speed of 25 posts per second for 640 × 480 size video images. Finally, the significance means a value of super pixel blocks is calculated in superpixel units, and the image is segmented into objects and backgrounds by combining with the Otsu threshold segmentation algorithm to mention the image. In this paper, the proposed algorithm experiments with several sets of Kor Kor resolution remote sensing images, respectively, and the Markov random field model and fully convolutional network (FCN) algorithm are used as the comparison algorithm. By comparing the experimental results qualitatively and quantitatively, it is shown that the algorithm in this paper has an obvious practical effect on contrast enhancement of digital media images and has certain practicality and superiority.

show abstract

“…NR-IQA methods can be divided into two categories: hand-crafted feature-based approaches and training-based approaches. The former methods [5][6][7][8][9][10][11][12][13] are widely applied to extract reliable features, which assume the authentic images share certain statistics and the emergence of distortions may change these statistics. However, since the methods rely on hand-crafted features designed for specific distortions, it is difficult to adequately represent the distortions of the distorted images.…”

Section: Introductionmentioning

confidence: 99%

No‐reference image quality assessment based on multiscale feature representation

Qiao

Zhao

et al. 2021

IET Image Processing

View full text Add to dashboard Cite

The no-reference image quality assessment (NR-IQA) method can evaluate the distortions in an image without the reference image. However, due to the diversity of the image contents and distortion types, it is hard for the existing NR-IQA algorithms to obtain competitive performance on both synthetically and authentically distorted images. To address the problem, a multiscale feature representation-based NR-IQA method that performs well for both synthetic and authentic distortions is proposed. This model consists of two parts: The feature extraction part and the feature fusion part. First, part of the Res2Net-50 network is chosen as the feature extraction part due to its high ability in increasing the range of receptive fields. Then, the feature fusion part consisting of a novel residual block and two fully connected layers is designed to fuse the extracted features and realize the quality score mapping. After a series of stepwise optimization experiments, the most competitive network architecture consisting of the feature extraction part and the feature fusion part is obtained. Comprehensive experiments on the LIVE, TID2013, CSIQ, KADID-10k, KonIQ-10K, and LIVE challenge databases demonstrate that the proposed method can work powerfully on both the synthetic and authentic distortions and also has a strong generalization ability.This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.

show abstract

Image quality assessment via spatial‐transformed domains multi‐feature fusion

Cited by 5 publications

References 40 publications

CPDINet: Blind image quality assessment via a content perception and distortion inference network

CPDINet: Blind image quality assessment via a content perception and distortion inference network

Multifeature Contrast Enhancement Algorithm for Digital Media Images Based on the Diffusion Equation

No‐reference image quality assessment based on multiscale feature representation

Contact Info

Product

Resources

About