Search

Blind image quality assessment (BIQA) is a fundamental yet challenging problem in the image processing system. Existing BIQA models have the following problems: 1) Due to the lack of available quality-label images, most of the methods have poor generalization ability in different distortion categories; 2) The impact of human visual characteristics on the content of images is not taken into account. In this paper, we proposed a visual residual perception optimized network (VRPON) that can effectively solve these problems. The proposed method separates the training of BIQA into two stages: 1) a distortion degree identification network and 2) an image quality prediction network. In the first stage, the spatial and temporal features of image sequences are extracted by CNN and LSTM respectively, which are used to evaluate the degree of image distortion. And then the proposed model is learned to predict image patches' scores in the second stage with the outputs of the first stage. Finally, a pooling strategy that follows the human visual saliency is designed to evaluate the quality score of the whole image. Experimental results show that the proposed VRPON not only has better performance than state-of-the-art methods on synthetic distorted images (LIVE, TID2013, CSIQ), but also has better robustness for different authentic distortions (LIVE challenge).

The significant target in the image shows how much people pay attention to certain objects in the image when viewing the image. These significant targets will affect human perception of image quality when quality change occurs. But in recent years, most of the image quality perception research based on elsctroencephalography(EEG) were not considered of these significant targets. Therefore, this paper obtains the EEG signals of the subjects through an image rendering experiment for the presence or absence of significant targets, and compares the signal differences. The experimental results show that there is no significant effect on the image quality perception in the perceptible range. If there is a large degree of distortion, whether there is significant target has an important impact on the image quality perception.

The state-of-the-art image quality assessment (IQA) methods based on structure calculation are mostly implemented without considering visual color sensitivity and visual perception threshold (VPT). Actually, human visual system (HVS) is extremely sensitive to the color structure and has visual perception threshold to reduce redundant information by representing the image sparsely. Hence, through incorporating the color structure and the VPT, this paper proposed a novel IQA framework which utilizes minimum amount of structure coefficients to capture the variation of color structure and distortion of degraded image by applying a VPT to remove the visual unperceived coefficients. The difference of the proportion of visual perceived coefficients between distorted and reference image is measured to acquire image quality score. Experimental results on two benchmark databases (LIVE II and TID) indicate the rationality and validity of the proposed method.

No-reference/Blind image quality assessment (NR-IQA/BIQA) is significant for image processing and yet very challenging, especially for real-time application and big image data processing. Traditional NR-IQA metrics usually train complex models such as support vector machine, neural network, and probability graph model, which result in long computational time and poor robustness. To overcome these weaknesses, the paper proposes a fast no-reference image quality assessment via hash coding method, named NRHC. First, the image is divided into overlapped patches to extract the spatial statistical features of natural scene images. Then the features are encoded to produce binary hash codes via supervised iterative quantization (SITQ) method. Finally, the Hamming distances between the hash code of the test image and those of original undistorted images are calculated to obtain the final image quality. Thorough experiments on benchmark databases demonstrate that the proposed approach achieves comparable performance and has higher computational efficiency and stronger robustness compared with the state-of-the-art NR-IQA methods.

The state-of-the-art blind image quality assessment (BIQA) metrics usually require a large amount of human scored images to train a regression model used to judge image quality, which makes the results are heavily dependent on the size of training data. In this paper, we present an efficient BIQA algorithm based on shearlet transform without using human scored images. This is mainly based on that the degradation of the image leads to significant variation in the spread discontinuities in all directions. However, shearlet transform has a strong ability to localize distributed discontinuities. The natural scene statistics (NSS) of shearlet coefficients are applicable to indicate the variation of image quality. Experimental results on benchmark databases illustrate that the proposed method has a good consistency with the subjective assessment of human beings.

Image Distortion Identification is important for image processing system enhancement, image distortion correction and image quality assessment. Although images may suffer various number of distortions while going through different systems, most of the previous researches of image distortion identification were focus on identifying single distortion in image. In this paper, we proposed a CNN-based multi-label learning model (called MLLNet) to identify distortions for different scenarios, including images having no distortion, single distortion and multiple distortions. Concretely, we transform the multi-label classification for image distortion identification to a number of multi-class classifications and use a deep multi-task CNN model to train all associated classifiers simultaneously. For unseen image, we use the trained CNN model to predict a number of classifications at same time and fuse them to final multi-label classification. The extensive experiments demonstrate that the propose algorithm can achieve good performance on several databases. Moreover, the network architecture of the CNN model can make flexible adjustment according to the different requirements.

Existing blind image quality assessment (BIQA) methods based on statistics attach limited attention to the relative position of pixels. Features in these BIQA methods are too flimsy to characterize quite a few distortions with strong locality or complexity. However, psychological studies have shown that according to the relative position within visual field, the cognitive system generates visuo-spatial serial memory used for cognitive tasks, e.g., subjective image quality assessment. Inspired by the visuo-spatial series generated by human visual system (HVS), we propose a BIQA method based on imitation Visuo-spatial Series Statistics (VSS). The proposed method simulates visual system to construct visuo-spatial series based on the relative position of pixels, and use statistical features of visuo-spatial series to predict image quality. Extensive experiments demonstrate the proposed method has a superior performance compared to the state-of-the-art BIQA methods.

Dehazed image quality assessment algorithm is aimed to evaluate the quality of dehazed images. However, the existing dehazed image quality evaluation algorithms are overly dependent on the dehazed images database with accurate subjective quality scores, which are inaccurate, biased and time-consuming, and it is difficult to obtain a large dehazed images database, or extent the existing database. To overcome these problems, it is by using the subjective quality preference that we propose a rank learning algorithm to evaluate the dehazed image quality, here, the subjective quality preference stands for the information such as “the quality of Image \(I_a\) is better than that of image \(I_b\)”, and the rank learning is aimed to learn a function that can predict the corresponding rank sequence for a given set of input stimulus. In our algorithm, we transform the problem of dehazed image quality evaluation into the classification problem of quality preference learning and then use and random forest and pairwise comparison in turn to learn the function that can predict the corresponding quality rank sequence for a given set of dehazed images. The experimental results show that our algorithm is highly consistent with the subjective feeling of human eye and is superior to the traditional dehazed image quality evaluation algorithms. Moreover, our algorithm has a strong expansibility.

Our eyes receive the information of the images containing both the luminance information and chrominance information. However, the available blind image quality assessment (BIQA) criteria usually involve luminance information only. In this paper, we propose a novel efficient IQA metric via statistics of color descriptors in the DCT domain. Firstly, we calculate the saturation (S), hue (H), luminance (L) of the testing image simultaneously. Then the local DCT transform is implemented on each color descriptor, and the nature scene statistics (NSS) are extracted from the DCT coefficients. This is mainly based on the fact that the degradation of the image induces considerable deviation in the frequency domain characteristics of chromatic data in natural image. However, the deviation can be quantified by the DCT coefficients of the image’s color descriptors effectively. Finally, we construct the mapping relation between the features and the image quality. Experimental results on several benchmarking databases (TID2013, LIVEII and CSIQ) show the proposed method is superior to other state-of-the-arts methods and reveal the rationality and the validity of the new approach.

Image quality assessment (IQA) is critically important for the image-processing field. IQA aims to build a computational model to predict human perceived image quality, accurately and automatically. Until now, great efforts have been employed to design IQA metrics. In this paper, we systematically and comprehensively review the fundamental, brief history, and state-of-the-art developments of IQA, with emphasis on natural image quality assessment (NIQA). First, the definition of image quality is discussed, which contains three aspects and lead to different philosophies of designing IQA metrics. Afterwards, classic NIQA metrics are presented with some further discussions. Widely used databases and the performances of classic NIQA metrics on them are also listed. We highlight the most significant works and some open issues about the developments of IQA, and provide the benchmarks for the researchers and scholars who work on IQA.

Considering that the Human Visual System (HVS) has different perceptual characteristics for different morphological components, a novel image quality metric is proposed by incorporating Morphological Component Analysis (MCA) and HVS, which is capable of assessing the image with different kinds of distortion. Firstly, reference and distorted images are decomposed into linearly combined texture and cartoon components by MCA respectively. Then these components are turned into perceptual features by Just Noticeable Difference (JND) which integrates masking features, luminance adaptation and Contrast Sensitive Function (CSF). Finally, the discrimination between reference and distorted images perceptual features is quantified using a pooling strategy before the final image quality is obtained. Experimental results demonstrate that the performance of the proposed prevails over some existing methods on LIVE database II. © 2011 World Scientific Publishing Company.

As the performance indicator of the image processing algorithms or systems, image quality assessment (IQA) has attracted the attention of many researchers. Aiming to the widely used compression standards, JPEG and JPEG2000, we propose a new no reference (NR) metric for compressed images to do IQA. This metric exploits the causes of distortion by JPEG and JPEG2000, employs the directional discrete cosine transform (DDCT) to obtain the detail and directional information of the images and incorporates with the visual perception to obtain the image quality index. Experimental results show that the proposed metric not only has outstanding performance on JPEG and JPEG2000 images, but also applicable to other types of artifacts.

The available image quality assessment (IQA) methods based on gradient calculation are mostly implemented without considering visual perception threshold (VPT) and color information. However, incorporating VPT with IQA model can reduce redundant information and human visual system (HVS) is extremely sensitive to color variation. An improved image quality assessment in gradient domain is proposed which utilizes minimum amount of gradient coefficients to capture the color and structure distortion of degraded image by applying a VPT to remove the unperceived gradient coefficients. The difference of perceived gradient coefficients between distorted and reference image is measured to acquire image quality score. Experimental results on two benchmarking databases (LIVEII and TID2008) indicate the rationality and validity of the proposed method.

Most of the existing assessment methods for color image quality consider little about the image content, which plays an essential role in indicating the distortion level with the color and structure information it contains. By incorporating the color image content with human color perception, we present a novel assessment metric for color distortions. The proposed method standardizes input images with a color perception transformation, and then the transformed images are divided into lightness part and chroma part. For each part, a region separation strategy based on image content is implemented. By calculating and pooling the similarity of each region using fuzzy integral, the final index is achieved. Experimental results on color-related distortions of TID2013 database show the superiority of this new approach and comparative experiments reveal the rationality and the validity of our method.

The available blind image quality assessment (BIQA) criteria usually involve a large amount of human scored images to train a regression model used to judge image quality, which makes the results are closely related to the amount of training data. In this paper, a valid BIQA algorithm based on shearlet transform without using human scored images is presented. This is mainly based on that degradation of image induces considerable deviation in the distributed discontinuities in different directions. However, the distributed discontinuities can be localized by shearlet transform effectively. The nature scene statistics (NSS) of shearlet coefficients are capable for exhibiting the alteration of image quality. Experimental results on two benchmarking databases (LIVEII and TID2008) indicate the rationality and validity of the proposed method.

The available image quality assessment methods are mostly based on statistical characteristic and consider very little the change of pixel correlation in conjunction with the quality assessment, which induces the quality assessment metric to be limited in the degradation of image quality caused by the change of pixel correlation. However, the pixel correlation change has a big effect on the image quality, so a novel image quality assessment based on the pixel correlation is proposed in this paper. Firstly image is parted based on mutual information, and then, three kinds of mutual information between the pixel intensity and the image patches are extracted to catch the variation of the pixel correlation. Finally the machine learning is utilized to learn the mapping from these differences space to image quality. The experimental results show that the proposed framework has good consistency with subjective perception.

Developing reduced reference image quality assessment (RR-IQA) plays a vital role in dealing with the prediction of the visual quality of distorted images. However, most of existing methods fail to take color information into consideration, although the color distortion is significant for the increasing color images. To solve the aforementioned problem, this paper proposed a novel IQA method which focuses on the color distortion. In particular, we extract color features based on the model of color fractal structure. Then the color and structure features are mapped into visual quality using the support vector regression. Experimental results on the LIVE II database demonstrate that the proposed method has a good consistency with the human perception especially on images with color distortion. © 2012 Springer-Verlag.

Image quality assessment (IQA) algorithms are important for image-processing systems. And structure information plays a significant role in the development of IQA metrics. In contrast to existing structure driven IQA algorithms that measure the structure information using the normalized image or gradient amplitudes, we present a new Local Structure Divergence (LSD) index based on the local structures contained in an image. In particular, we exploit the steering kernels to describe local structures. Afterward, we estimate the quality of a given image by calculating the symmetric Kullback-Leibler divergence (SKLD) between kernels of the reference image and the distorted image. Experimental results on the LIVE database II show that LSD performs consistently with the human perception with a high confidence, and outperforms representative structure driven IQA metrics across various distortions. © 2012 Springer-Verlag.

Blind image quality assessment (BIQA) is an important yet difficult task in image processing related applications. Existing algorithms for universal BIQA learn a mapping from features of an image to the corresponding subjective quality or divide the image into different distortions before mapping. Although these algorithms are promising, they face the following problems: 1) they require a large number of samples (pairs of distorted image and its subjective quality) to train a robust mapping; 2) they are sensitive to different datasets; and 3) they have to be retrained when new training samples are available. In this paper, we introduce a simple yet effective algorithm based upon the sparse representation of natural scene statistics (NSS) feature. It consists of three key steps: extracting NSS features in the wavelet domain, representing features via sparse coding, and weighting differential mean opinion scores by the sparse coding coefficients to obtain the final visual quality values. Thorough experiments on standard databases show that the proposed algorithm outperforms representative BIQA algorithms and some full-reference metrics. © 2012 IEEE.

No reference image quality assessment (NR-IQA) is to evaluate image quality blindly without the ground truth. Most of the emerging NR-IQA algorithms are only effective for some specific distortion. Universal metrics that can work for various categories of distortions have hardly been explored, and the algorithms available are not fully adequate in performance. In this paper, we study the local dependency (LD) characteristic of natural images, and propose two universal NR-IQA metrics: LD global scheme (LD-GS) and LD two-step scheme (LD-TS). We claim that the local dependency characteristic among wavelet coefficients is disturbed by various distortion processes, and the disturbances are strongly correlated to image qualities. Experimental results on LIVE database II demonstrate that both the proposed metrics are highly consistent with the human perception and outpace the state-of-the-art NR-IQA indexes and some full reference quality indicators for diverse distortions and across the entire database. © 2011 IEEE.

Due to that human eye has different perceptual characteristics for different morphological components, so a novel image quality metric is proposed by incorporating morphological component analysis (MCA) and human visual system (HVS), which is capable of assessing the image with different types of distortion. Firstly, reference and distorted images are decomposed into texture and cartoon components by MCA respectively. Then these components are changed into perceptual features by just noticeable difference (JND) which integrates masking features, luminance adaptation and contrast sensitive function (CSF). Finally, the difference between reference and distorted images' perceptual features is quantified using a pooling strategy, and then the final result of the image quality is obtained. Experimental results demonstrate that the performance of the metric prevail over some existing methods on LIVE database II. ©2010 IEEE.

Blind image quality assessment (BIQA) is a challenging task due to the difficulties in extracting quality-aware features and modeling the relationship between the image features and the visual quality. Until now, most BIQA metrics available try to extract statistical features based on the natural scene statistics (NSS) and build mapping from the features to the quality score using the supervised machine learning technique based on a large amount of labeled images. Although several promising metrics have been proposed based on the above methodology, there are two drawbacks of these algorithms. First, only the labeled images are adopted for machine learning. However, it has been proved that using unlabeled data in the training stage can improve the learning performance. In addition, these metrics try to learn a direct mapping from the features to the quality score. However, subjective quality evaluation would be rather a fuzzy process than a distinctive one. Equally, human beings tend to evaluate the quality of a given image by first judging the extents it belongs to “excellent,” “good,” “fair,” “bad,” and “poor,” and estimating the quality score subsequently, rather than directly giving an exact subjective quality score. To overcome the aforementioned problems, we propose a semi-supervised and fuzzy framework for blind image quality assessment, S2F2, in this paper. In the proposed framework, (1) we formulate the fuzzy process of subjective quality assessment by using fuzzy inference. Specially, we model the membership relation between the subjective quality score and the truth values it belongs to “excellent,” “good,” “fair,” “bad,” and “poor” using a Gaussian function, respectively; and (2) we introduce the semi-supervised local linear embedding (SS-LLE) to learn the mapping function from the image features to the truth values using both the labeled and unlabeled images. In addition, we extract image features based on NSS since it has led to promising performances for image quality assessment. Experimental results on two benchmarking databases, i.e., the LIVE database II and the TID2008 database, demonstrate the effectiveness and promising performance of the proposed S2F2 algorithm for BIQA.