Search

The paper describes a new image database HTID for verification and training of no-reference image visual quality metrics. The database contains 2880 color images of size 1536×1024 pixels cropped from the real-life photos produced by the mobile phone cameras with various shooting and post-processing settings. Mean opinion scores for images of the database are obtained. Peculiarities of the database are considered. A comparative analysis of the state-of-The-Art no-reference image visual quality metrics is carried out. It is shown that the proposed database takes its own unique place in the existing image databases and can be effectively used for metrics' verification. acceptedVersion

Remote sensing images are subject to different types of degradations. The visual quality of such images is important because their visual inspection and analysis are still widely used in practice. To characterize the visual quality of remote sensing images, the use of specialized visual quality metrics is desired. Although the attempts to create such metrics are limited, there is a great number of visual quality metrics designed for other applications. Our idea is that some of these metrics can be employed in remote sensing under the condition that those metrics have been designed for the same distortion types. Thus, image databases that contain images with types of distortions that are of interest should be looked for. It has been checked what known visual quality metrics perform well for images with such degradations and an opportunity to design neural network-based combined metrics with improved performance has been studied. It is shown that for such combined metrics, their Spearman correlation coefficient with mean opinion score exceeds 0.97 for subsets of images in the Tampere Image Database (TID2013). Since different types of elementary metric pre-processing and neural network design have been considered, it has been demonstrated that it is enough to have two hidden layers and about twenty inputs. Examples of using known and designed visual quality metrics in remote sensing are presented. publishedVersion

International audience; This paper deals with analysis of component-wise and three-dimensional (3D) lossy compression of hyperspectral data. The latter approach is known to be more efficient due to possibility to employ inter-band correlation of images in neighbor sub-bands. However, there are certain peculiarities for this approach which are subject of our study. Methods of lossy compression that employ discrete cosine transform (DCT) in blocks are studied. It is shown that 3D compression is able to provide considerable benefits in compression ratio and/or mean square error (MSE) of introduced distortions but only under certain conditions. Experiments have been done for real-life hyperspectral data with high input signal-to-noise ratio. Their main results are presented. © 2019 IEEE.

This paper is devoted to assessing effectiveness of verification of no-reference quality metrics. The analysis of the impact of image visual quality on accuracy of verification of quality metrics is carried out. The problem of estimating metrics adequacy for images with almost unnoticeable distortions is shown. Modifications of rank correlation coefficients are proposed to address this problem. The results are presented for several good particular quality metrics as well as the designed neural network based combined visual quality metric.

A task of time delay estimation using two or more sensors that receive and process noise-like wideband signals embedded in spatially uncorrelated noise can be treated as classical. A problem arises if environment noise is non-Gaussian and has heavy tails. Then, due to several factors, probability of appearing abnormal elementary time delay estimates can be large, especially if conventional method of cross-correlation signal processing is applied. Different methods to cope with the problem have been proposed including the use of robust DFT at initial stages. However, this reduces computational efficiency of processing radically. Here, we propose a specific modification of classical processing that presumes pre-processing of noisy received signals using center weighted median filter. We show that this easy and fast step allows providing better performance than procedures based on robust DFT. Considerable reduction of abnormal error probability is observed and the method can operate in conditions of limited a priori information about characteristics of additive non-Gaussian noise.

The problem of increasing efficiency of blind image quality assessment is considered. No-reference image quality metrics both independently and as components of complex image processing systems are employed in various application areas where images are the main carriers of information. Meanwhile, existing no-reference metrics have a significant drawback characterized by a low adequacy to image perception by human visual system (HVS). Many well-known no-reference metrics are analyzed in our paper for several image databases. A method of combining several no-reference metrics based on artificial neural networks is proposed based on multi-database verification approach. The effectiveness of the proposed approach is confirmed by extensive experiments. publishedVersion

This paper studies the problem of full reference visual quality assessment of denoised images with a special emphasis on images with low contrast and noise-like texture. Denoising of such images together with noise removal often results in image details loss or smoothing. A new test image database, FLT, containing 75 noise-free "reference" images and 300 filtered ("distorted") images is developed. Each reference image, corrupted by an additive white Gaussian noise, is denoised by the BM3D filter with four different values of threshold parameter (four levels of noise suppression). After carrying out a perceptual quality assessment of distorted images, the mean opinion scores (MOS) are obtained and compared with the values of known full reference quality metrics. As a result, the Spearman Rank Order Correlation Coefficient (SROCC) between PSNR values and MOS has a value close to zero, and SROCC between values of known full-reference image visual quality metrics and MOS does not exceed 0.82 (which is reached by a new visual quality metric proposed in this paper). The FLT dataset is more complex than earlier datasets used for assessment of visual quality for image denoising. Thus, it can be effectively used to design new image visual quality metrics for image denoising., Comment: Submitted to ICASSP 2018

This paper describes a recently created image database, TID2013, intended for evaluation of full-reference visual quality assessment metrics. With respect to TID2008, the new database contains a larger number (3000) of test images obtained from 25 reference images, 24 types of distortions for each reference image, and 5 levels for each type of distortion. Motivations for introducing 7 new types of distortions and one additional level of distortions are given; examples of distorted images are presented. Mean opinion scores (MOS) for the new database have been collected by performing 985 subjective experiments with volunteers (observers) from five countries (Finland, France, Italy, Ukraine, and USA). The availability of MOS allows the use of the designed database as a fundamental tool for assessing the effectiveness of visual quality. Furthermore, existing visual quality metrics have been tested with the proposed database and the collected results have been analyzed using rank order correlation coefficients between MOS and considered metrics. These correlation indices have been obtained both considering the full set of distorted images and specific image subsets, for highlighting advantages and drawbacks of existing, state of the art, quality metrics. Approaches to thorough performance analysis for a given metric are presented to detect practical situations or distortion types for which this metric is not adequate enough to human perception. The created image database and the collected MOS values are freely available for downloading and utilization for scientific purposes. We have created a new large database.This database contains larger number of distorted images and distortion types.MOS values for all images are obtained and provided.Analysis of correlation between MOS and a wide set of existing metrics is carried out.Methodology for determining drawbacks of existing visual quality metrics is described.

Specialized image databases like TID2013, allow quantitative evaluation of adequateness to human perception characterized by mean opinion score (MOS) and a verified full-reference visual quality metric using certain criterions such as rank order correlation coefficients (ROCC). In this paper, we propose modifications of known Spearman and Kendall ROCCs that take into account the fact that MOS for existing databases has been measured with a limited accuracy and is most certainly erroneous (contains “noise component”). For the database TID2013 we have also marked images with practically invisible distortions. This has led to considering criteria for detecting such images. Using a mean square error we also analyze metrics based on alternative norms l1, l3, and l4. Analysis is performed for the database TID2013 and different sensitivity of human vision system (HVS) to distortions in color and intensity is taken into account. One approach to analyze adequateness of metrics for different types of distortions is described and proved exploitable.