A Regression-Based Family of Measures for Full-Reference Image Quality Assessment

Abstract The advances in the development of imaging devices resulted in the need of an automatic quality evaluation of displayed visual content in a way that is consistent with human visual perception. In this paper, an approach to full-reference image quality assessment (IQA) is proposed, in which several IQA measures, representing different approaches to modelling human visual perception, are efficiently combined in order to produce objective quality evaluation of examined images, which is highly correlated with evaluation provided by human subjects. In the paper, an optimisation problem of selection of several IQA measures for creating a regression-based IQA hybrid measure, or a multimeasure, is defined and solved using a genetic algorithm. Experimental evaluation on four largest IQA benchmarks reveals that the multimeasures obtained using the proposed approach outperform state-of-the-art full-reference IQA techniques, including other recently developed fusion approaches.

Download Full-text

A Novel Full Reference-Image Quality Assessment (FR-IQA) for Adaptive Visual Perception Improvement

2021 6th International Conference on Inventive Computation Technologies (ICICT) ◽

10.1109/icict50816.2021.9358610 ◽

2021 ◽

Author(s):

D Narsaiah ◽

R Surender Reddy ◽

Aruna Kokkula ◽

P Anil Kumar ◽

A Karthik

Keyword(s):

Image Quality ◽

Visual Perception ◽

Quality Assessment ◽

Image Quality Assessment ◽

Reference Image ◽

Full Reference

Download Full-text

Quality Assessment on Authentically Distorted Images by Expanding Proxy Labels

Electronics ◽

10.3390/electronics9020252 ◽

2020 ◽

Vol 9 (2) ◽

pp. 252 ◽

Cited By ~ 1

Author(s):

Xiaodi Guan ◽

Fan Li ◽

Lijun He

Keyword(s):

Image Quality ◽

Visual Perception ◽

Quality Assessment ◽

Image Quality Assessment ◽

Image Features ◽

Quality Score ◽

Fine Tuning ◽

Reference Image ◽

Human Visual Perception ◽

In The Wild

In this paper, we propose a no-reference image quality assessment (NR-IQA) approach towards authentically distorted images, based on expanding proxy labels. In order to distinguish from the human labels, we define the quality score, which is generated by using a traditional NR-IQA algorithm, as “proxy labels”. “Proxy” means that the objective results are obtained by computer after the extraction and assessment of the image features, instead of human judging. To solve the problem of limited image quality assessment (IQA) dataset size, we adopt a cascading transfer-learning method. First, we obtain large numbers of proxy labels which denote the quality score of authentically distorted images by using a traditional no-reference IQA method. Then the deep network is trained by the proxy labels, in order to learn IQA-related knowledge from the amounts of images with their scores. Ultimately, we use fine-tuning to inherit knowledge represented in the trained network. During the procedure, the mapping relationship fits in with human visual perception closer. The experimental results demonstrate that the proposed algorithm shows an outstanding performance as compared with the existing algorithms. On the LIVE In the Wild Image Quality Challenge database and KonIQ-10k database (two standard databases for authentically distorted image quality assessment), the algorithm realized good consistency between human visual perception and the predicted quality score of authentically distorted images.

Download Full-text

Combined Full-Reference Image Quality Metrics for Objective Assessment of Multiply Distorted Images

Electronics ◽

10.3390/electronics10182256 ◽

2021 ◽

Vol 10 (18) ◽

pp. 2256

Author(s):

Krzysztof Okarma ◽

Piotr Lech ◽

Vladimir V. Lukin

Keyword(s):

Image Quality ◽

Quality Assessment ◽

Image Quality Assessment ◽

Objective Assessment ◽

Assessment Methods ◽

Subjective Quality ◽

Reference Image ◽

Exponential Weights ◽

Full Reference ◽

Highly Correlated

In the recent years, many objective image quality assessment methods have been proposed by different researchers, leading to a significant increase in their correlation with subjective quality evaluations. Although many recently proposed image quality assessment methods, particularly full-reference metrics, are in some cases highly correlated with the perception of individual distortions, there is still a need for their verification and adjustment for the case when images are affected by multiple distortions. Since one of the possible approaches is the application of combined metrics, their analysis and optimization are discussed in this paper. Two approaches to metrics’ combination have been analyzed that are based on the weighted product and the proposed weighted sum with additional exponential weights. The validation of the proposed approach, carried out using four currently available image datasets, containing multiply distorted images together with the gathered subjective quality scores, indicates a meaningful increase of correlations of the optimized combined metrics with subjective opinions for all datasets.

Download Full-text

Adaptation of Full-Reference Image Quality Assessment Methods for Automatic Visual Evaluation of the Surface Quality of 3D Prints

Elektronika ir Elektrotechnika ◽

10.5755/j01.eie.25.5.24357 ◽

2019 ◽

Vol 25 (5) ◽

pp. 57-62 ◽

Cited By ~ 4

Author(s):

Krzysztof Okarma ◽

Jaroslaw Fastowicz

Keyword(s):

Additive Manufacturing ◽

Image Quality ◽

Quality Assessment ◽

Surface Quality ◽

Quality Evaluation ◽

Image Quality Assessment ◽

Assessment Methods ◽

Reference Image ◽

Visual Evaluation ◽

Full Reference

Automatic visual quality assessment of the 3D printed surfaces is currently one of the most demanding challenges in additive manufacturing. Regardless of the applications of the computer vision for the 3D printing process monitoring purposes, a reliable surface quality evaluation during manufacturing may introduce brand new possibilities. The detection of some distortions and their automatic evaluation can be helpful when deciding to stop the process to save time, energy, and filament. In some cases, some further corrections can also be made for relatively small distortions. Since many general-purpose image quality assessment methods have been proposed in recent years, their applications for the quality evaluation in the additive manufacturing are investigated. As most of the metrics are full-reference and require the availability of the original perfect quality image, their direct application is not possible. Therefore, their adaptation is described in the paper together with experimental verification of classification results obtained using various metrics.

Download Full-text

A comprehensive assessment of content-based image retrieval using selected full reference image quality assessment algorithms

Multimedia Tools and Applications ◽

10.1007/s11042-021-10573-0 ◽

2021 ◽

Author(s):

Baisakhi Sur Phadikar ◽

Amit Phadikar ◽

Subro S. Thakur

Keyword(s):

Image Quality ◽

Image Retrieval ◽

Quality Assessment ◽

Image Quality Assessment ◽

Comprehensive Assessment ◽

Content Based Image Retrieval ◽

Reference Image ◽

Full Reference

Download Full-text

Precise No-Reference Image Quality Evaluation Based on Distortion Identification

ACM Transactions on Multimedia Computing Communications and Applications ◽

10.1145/3468872 ◽

2021 ◽

Vol 17 (3s) ◽

pp. 1-21

Author(s):

Chenggang Yan ◽

Tong Teng ◽

Yutao Liu ◽

Yongbing Zhang ◽

Haoqian Wang ◽

...

Keyword(s):

Neural Network ◽

Image Quality ◽

Quality Assessment ◽

Large Scale ◽

Quality Evaluation ◽

Image Quality Assessment ◽

State Of The Art ◽

Gaussian White Noise ◽

The State ◽

Reference Image

The difficulty of no-reference image quality assessment (NR IQA) often lies in the lack of knowledge about the distortion in the image, which makes quality assessment blind and thus inefficient. To tackle such issue, in this article, we propose a novel scheme for precise NR IQA, which includes two successive steps, i.e., distortion identification and targeted quality evaluation. In the first step, we employ the well-known Inception-ResNet-v2 neural network to train a classifier that classifies the possible distortion in the image into the four most common distortion types, i.e., Gaussian white noise (WN), Gaussian blur (GB), jpeg compression (JPEG), and jpeg2000 compression (JP2K). Specifically, the deep neural network is trained on the large-scale Waterloo Exploration database, which ensures the robustness and high performance of distortion classification. In the second step, after determining the distortion type of the image, we then design a specific approach to quantify the image distortion level, which can estimate the image quality specially and more precisely. Extensive experiments performed on LIVE, TID2013, CSIQ, and Waterloo Exploration databases demonstrate that (1) the accuracy of our distortion classification is higher than that of the state-of-the-art distortion classification methods, and (2) the proposed NR IQA method outperforms the state-of-the-art NR IQA methods in quantifying the image quality.

Download Full-text