Mean Opinion Score Values Research Articles

Xception-based architecture with cross-sampled training for Image Quality Assessment on KonIQ-10k

Image quality assessment is a~crucial task in various fields such as digital photography, online content creation, and automated quality control, as it ensures an optimal visual experience and aids in maintaining consistent standards. In this paper, we propose an efficient method for training image quality assessment models on the KonIQ-10k dataset. Our novel approach utilizes a~dual-Xception architecture that analyzes both the image content and additional image parameters, outperforming traditional single convolutional models. We introduce cross-sampling methods with random draw sampling of instances from majority classes, effectively enhancing prediction quality in the Mean Opinion Score (MOS) ranges that are underrepresented in the database. This methodology allows us to achieve near state-of-the-art results with limited computing costs and resources. Most importantly, our predictions across the entire spectrum of MOS values maintain consistent quality. Because of using a~novel and highly effective method for image sampling, we achieved these results with much lower computational cost, making our approach the most effective way of MOS estimation on the KonIQ-10k database.

QUALITY OF EXPERIENCE FOR VOICE OVER INTERNET PROTOCOL (VoIP)

Today, the service of Voice over Internet Protocol or (VoIP) is one of the most used service around the globe in many fields, especially with video conferencing applications. Different applications and also many communications companies around the world use the VoIP service. Communication stakeholders aim to keep old customers happy and attract new customers. There was a need for some kind of service quality measures. Metrics are used to quantify the quality of transmission link are known Quality of Service (QoS). Whereas, user satisfaction level is identified by using Quality of Experience (QoE). This paper conducted a simulation model for VoIP service over heterogeneous, using OPNET modeler. This work aims to compare the user satisfaction level of the VoIP service by using different coding schemes in the application level of the User Equipment (UE). The measured QoS parameters over the heterogeneous transmission link were IP packet delay, IP Packet jitter, IP packet loss, MAC layer delay, PHY layer throughput in the network. Additionally, in this simulation model the use of E-Model to assess QoE level using Mean Opinion Score (MOS) metric in predicting the VoIP call quality. The results using the vocoder G.729A during the (30) sec of the voice call simulation time were as follows: Delay 0.57 mSec, Throughput of 290 Kbits/sec, Jitter of 0.02 uSec, MOS value of 3.76. The results using the vocoder G.711 during the same aforementioned time were as follows: Delay 0.25 mSec, Throughput of 70 Kbits/sec, Jitter of 0 Sec), MOS value of 3.08.

Subjective Quality Assessment of V-PCC-Compressed Dynamic Point Clouds Degraded by Packet Losses.

This article describes an empirical exploration on the effect of information loss affecting compressed representations of dynamic point clouds on the subjective quality of the reconstructed point clouds. The study involved compressing a set of test dynamic point clouds using the MPEG V-PCC (Video-based Point Cloud Compression) codec at 5 different levels of compression and applying simulated packet losses with three packet loss rates (0.5%, 1% and 2%) to the V-PCC sub-bitstreams prior to decoding and reconstructing the dynamic point clouds. The recovered dynamic point clouds qualities were then assessed by human observers in experiments conducted at two research laboratories in Croatia and Portugal, to collect MOS (Mean Opinion Score) values. These scores were subject to a set of statistical analyses to measure the degree of correlation of the data from the two laboratories, as well as the degree of correlation between the MOS values and a selection of objective quality measures, while taking into account compression level and packet loss rates. The subjective quality measures considered, all of the full-reference type, included point cloud specific measures, as well as others adapted from image and video quality measures. In the case of image-based quality measures, FSIM (Feature Similarity index), MSE (Mean Squared Error), and SSIM (Structural Similarity index) yielded the highest correlation with subjective scores in both laboratories, while PCQM (Point Cloud Quality Metric) showed the highest correlation among all point cloud-specific objective measures. The study showed that even 0.5% packet loss rates reduce the decoded point clouds subjective quality by more than 1 to 1.5 MOS scale units, pointing out the need to adequately protect the bitstreams against losses. The results also showed that the degradations in V-PCC occupancy and geometry sub-bitstreams have significantly higher (negative) impact on decoded point cloud subjective quality than degradations of the attribute sub-bitstream.

An Intelligent Handover Mechanism Based on MOS Predictions for Real-Time Video Conference Services in Mobile Networks

In mobile networks, handover mechanisms provide fast and smooth access service for mobile users. However, one of the main challenges in mobile networks is the handover management with increased mobility and bandwidth demand of the required network services. Therefore, in this paper, we propose a MOS-aware (mean opinion score-aware) mobile network handover mechanism based on deep learning to determine the appropriate handover time for real-time video conference services in mobile networks. We construct a wireless network topology with LTE characteristics in a Mininet-WiFi simulation. User equipment (UE) can determine the service-required MOS (Mean Opinion Score) from the proposed deep-learning-based handover mechanism with appropriate handover time. Simulation results show that the proposed scheme provides higher performance than the original A3 handover mechanism. The contribution of this paper is to combine the real-time video conferencing services with a deep-learning-based handover mechanism by predicting MOS values to improve the quality of service for users in mobile networks.

A comparison of non-experts and experts using DSIS method

Subjective evaluations are necessary to learn how expected viewers perceive the quality of a system. Traditionally, non-expert subjective tests are preferred rather than expert tests. In this study, we conducted subjective evaluation experiments for non-experts and experts on compressed 8K videos using the double stimulus impairment scale (DSIS) method and analyzed the experimental results expressed in terms of the mean opinion score (MOS), which is the average of individual scores. Furthermore, we investigated the differences between non-experts and experts by considering a new method in P.913 that estimates an improved MOS and a new experimental method using experts, called expert viewing protocol (EVP). Our contribution shows advantages of conducting expert subjective tests, such as EVP: expert tests allow to perform experiments with fewer subjects, to distinguish between original and distorted images, to determine a lower threshold for the image quality, to distribute scores in an appropriate range, and to constantly gain MOS values equal to improved MOS values.

Modeling Perceived Quality on 8K VVC Video Under Various Screen Sizes and Viewing Distances

Perceptual video quality considerably affects the quality of experience (QoE) of watching television (TV) broadcasts. Viewing conditions, such as the screen size and viewing distance, impact the perceived quality. We performed subjective evaluation experiments on 8K (7,680×4,320) ultra-high definition (UHD) compressed videos under seven viewing conditions (combinations of 31.5-, 55-, and 85-inch displays and 0.75, 1.5, and 3.0 H (times of screen height) of viewing distance). Distorted videos compressed by the versatile video coding (VVC)/H.266 were used in four types of encoding resolution, from 2K (1,920×1,080) to 8K, at a wide bitrate settings range of 3–80 Mbps. We derived a simple regression equation predicting the mean opinion score (MOS) using the hierarchical linear model (HLM), investigating the factors influencing subjective video quality. In this equation, MOS is expressed as a linear combination of terms including intercept and bitrate associated with sequence and encoding resolution, screen size, and viewing distance; it indicates that the smaller the screen, or the further the viewing distance, the fewer artifacts are perceived, as following empirical rules. Furthermore, we confirmed that the derived model is accurate as the Pearson linear and Spearman rank order correlation coefficients between predicted and actual MOS values were more than 0.97.

RM-IQA: A new no-reference image quality assessment framework based on range mapping method

Significant progress has been made in recent years in image quality assessment (IQA). In particular, the development of deep learning has provided no-reference (NR)-IQA with more impressive solutions. However, improving the generalization of NR-IQA models is still an urgent necessity. In this study, we propose a new framework that uses the range mapping method to map an existing full-reference (FR)-IQA dataset to an NR-IQA dataset, thereby further enhancing the accuracy and generalization of the NR-IQA model. First, an NR-IQA model is employed to score an FR-IQA dataset to obtain the corresponding mean opinion score (MOS) values. Then, the correlation coefficients between these MOS values and the original differential mean opinion score (DMOS) values marked by the FR-IQA dataset itself is calculated. Subsequently, the matching sequence pair is obtained according to these correlation coefficients. Then, a range mapping function is selected based on this sequence pair, and this function is used to map the entire FR-IQA dataset to the existing NR-IQA dataset, and a new NR-IQA dataset is generated. Finally, the new and the existing NR-IQA datasets are merged into a new dataset, which can train an end-to-end multi-task network to obtain the final model RM-IQA. This model exhibits better performance as it exploits more prior information. Based on the largest available NR-IQA dataset KonIQ-10k and FR-IQA dataset KADID-10K, the experimental results proved the effectiveness of the proposed framework.

Sistem Pakar Diagnosa Penyakit Kulit pada Manusia dengan Metode Dempster Shafer

Skin is the broadest organ in the human body that covers the entire surface of the body and acts as a support for human life. Because of its outermost location, the skin is often attacked by various diseases. This research aims to build an expert system to diagnose 10 types of skin diseases caused by viruses, fungi, bacteria, and parasites based on the knowledge of 3 experts using the Dempster Shafer method to obtain conclusions of skin diseases. Each symptom of skin disease has a value of belief that is used to calculate conclusions in the Dempster Shafer method. Expert system applications that are built can run Android-based smartphones. Testing techniques used in this research are black-box testing, theoretical calculations, system accuracy and MOS (Mean Opinion Score). Comparison of the calculation on the system has been appropriate based on the theoretical calculations. System accuracy testing in 30 sample cases resulted in an accuracy of 90%, but if it was seen as a subsection of expert diagnosis, it resulted in a system accuracy of 92.22%. MOS testing on 30 respondents results in a MOS value of 4.24 from a scale of 5 which shows that the system is proper to use and categorized into a good system.

Modeling and estimating the subjects\u2019 diversity of opinions in video quality assessment: a neural network based approach

Subjective experiments are considered the most reliable way to assess the perceived visual quality. However, observers’ opinions are characterized by large diversity: in fact, even the same observer is often not able to exactly repeat his first opinion when rating again a given stimulus. This makes the Mean Opinion Score (MOS) alone, in many cases, not sufficient to get accurate information about the perceived visual quality. To this aim, it is important to have a measure characterizing to what extent the observed or predicted MOS value is reliable and stable. For instance, the Standard deviation of the Opinions of the Subjects (SOS) could be considered as a measure of reliability when evaluating the quality subjectively. However, we are not aware of the existence of models or algorithms that allow to objectively predict how much diversity would be observed in subjects’ opinions in terms of SOS. In this work we observe, on the basis of a statistical analysis made on several subjective experiments, that the disagreement between the quality as measured by means of different objective video quality metrics (VQMs) can provide information on the diversity of the observers’ ratings on a given processed video sequence (PVS). In light of this observation we: i) propose and validate a model for the SOS observed in a subjective experiment; ii) design and train Neural Networks (NNs) that predict the average diversity that would be observed among the subjects’ ratings for a PVS starting from a set of VQMs values computed on such a PVS; iii) give insights into how the same NN based approach can be used to identify potential anomalies in the data collected in subjective experiments.

Effect of Motion Intensity on Perceived Quality of Multimedia Communication

Wireless multimedia communication facilitates the way we communicate and work. Multimedia communication has greatly changed the approach of modern world communication, especially during the peak period of coronavirus pandemic, where patterns of official meetings, business transactions and medical services shifted toward virtual approach using multimedia applications such as video conference, Skype, zoom applications and Video on Demand for personalized media consumption. Multimedia communication demands large chuck of scarced network resources to meet users’ quality performance compared to audio communication. This paper assesses the effect of motion intensity on perceived quality of multimedia communication. System simulations performed with the four different ITU-T reference sequences standard test multimedia sequences of various motion intensity characteristics shows that the perceived quality multimedia test sequences decreases with increase in motion intensity level of test multimedia samples under constraint network condition. Approximately, Akiyo test sample with significant low motion intensity recorded average Mean Opinion Score (MOS) value of 4.16 compared with 3.11 and 3.02 MOS values obtained for test samples with relative high motion intensity characteristics.

Expert System of Diagnosing Building Damage due to Earthquake using Backpropagation Artificial Neural Network Method

Earthquake is one of the most destructive natural disasters. After the earthquake, experts were deployed to survey the damage that occurred. One of the main objectives of the assessment task carried out by experts is to evaluate and classify buildings into several categories based on the level of damage that occurs. In this study, an expert system that could facilitate the assessment of building damage due to the earthquake was made using Backpropagation neural network method. The testing techniques used in this system are blackbox, accuracy and Mean Opinion Score (MOS) testing. MOS testing conducted by 30 respondents produced an MOS value of 4.54 from a scale of 5. While the average accuracy of the system obtained is 82.22% of the 30 case cases tested by 3 building damage experts.

No-reference quality assessment for image-based assessment of economically important tropical woods.

Image Quality Assessment (IQA) is essential for the accuracy of systems for automatic recognition of tree species for wood samples. In this study, a No-Reference IQA (NR-IQA), wood NR-IQA (WNR-IQA) metric was proposed to assess the quality of wood images. Support Vector Regression (SVR) was trained using Generalized Gaussian Distribution (GGD) and Asymmetric Generalized Gaussian Distribution (AGGD) features, which were measured for wood images. Meanwhile, the Mean Opinion Score (MOS) was obtained from the subjective evaluation. This was followed by a comparison between the proposed IQA metric, WNR-IQA, and three established NR-IQA metrics, namely Blind/Referenceless Image Spatial Quality Evaluator (BRISQUE), deepIQA, Deep Bilinear Convolutional Neural Networks (DB-CNN), and five Full Reference-IQA (FR-IQA) metrics known as MSSIM, SSIM, FSIM, IWSSIM, and GMSD. The proposed WNR-IQA metric, BRISQUE, deepIQA, DB-CNN, and FR-IQAs were then compared with MOS values to evaluate the performance of the automatic IQA metrics. As a result, the WNR-IQA metric exhibited a higher performance compared to BRISQUE, deepIQA, DB-CNN, and FR-IQA metrics. Highest quality images may not be routinely available due to logistic factors, such as dust, poor illumination, and hot environment present in the timber industry. Moreover, motion blur could occur due to the relative motion between the camera and the wood slice. Therefore, the advantage of WNR-IQA could be seen from its independency from a "perfect" reference image for the image quality evaluation.

Quality perception of specific chromatic impairments

This work was motivated by previous studies on the perceptual influence of chromatic impairments, where was observed that specific chromatic impairments might have a different influence, related to the represented content. Hence, in this study, chromatic impairments were applied to specific color clusters and the resulting perceptual influence was analyzed. The applied chromatic impairments influence on the images naturalness was assessed using a Single Stimulus Continuous Quality Evaluation (SSCQE) and the Mean Opinion Score (MOS) was calculated. Furthermore, during the image assessment, the eye movements were registered using an eye tracking device and the Relative Fixation Time (RFT) was computed. It was concluded that the induced chromatic impairments, lead to lower MOS values than the original images, revealing a perception of quality loss. Furthermore, when nature colors are changed, subjects revealed a larger perception of the impairment, resulting in smaller MOS and producing evident changes in the RFT’s. The comparative analysis between the MOS and the RFT’s reveals high Pearson Correlation Coefficient (PCC) results (p = 0.79 ± 0.14).

An expandable image database for evaluation of full-reference image visual quality metrics

Traditional approach to collect mean opinion score (MOS) values for evaluation of full-reference image quality metrics has two serious drawbacks. The first drawback is a nonlinearity of MOS, only partially compensated by the use of rank order correlation coefficients in a further analysis. The second drawback are limitations on number of distortion types and distortion levels in image database imposed by a maximum allowed time to carry out an experiment. One of the largest of databases used for this purpose, TID2013, has almost reached these limitations, which makes an extension of TID2013 within the boundaries of this approach to be practically unfeasible. In this paper, a novel methodology to collect MOS values, with a possibility to infinitely increase a size of a database by adding new types of distortions, is proposed. For the proposed methodology, MOS values are collected for pairs of distortions, one of them being a signal dependent Gaussian noise. A technique of effective linearization and normalization of MOS is described. Extensive experiments for linearization of MOS values to extend TID2013 database are carried out.

Blind quality assessment metric and degradation classification for degraded document images

Epochal documents suffer from several types of noises that accumulate and evolve over time. This significantly affects their quality and makes their storage and the interpretation of their visual content problematic. Digital preservation seems the most viable and the most promising. Moreover, measuring the amount of degradation and quality assessment of degraded documents is highly desirable for applications such as selecting the proper algorithms for enhancement and analysis of document images, filtering the damaged images, tuning the processing algorithms parameters, document repairing, psychological study, etc. The first contribution of this work is the proposition of an efficient Multi-distortion Document Quality Measure (MDQM) for quality assessment of physically degraded document images. The proposed MDQM metric is based on three sets of spatial and frequency image features. These features are extracted from two layers of text and non-text and mapped to the mean opinion scores (MOS) using the regression function. The second contribution of this work is to estimate the probability of four common document image distortion types, namely, paper translucency, stain, readers annotations and worn holes in the degraded images. In our experiment, the correlations of seven no-reference image quality assessment (NR-IQA) metrics with the MOS values are evaluated on two available datasets. It is shown that the performance of MDQM metric is significantly better than the state-of-the-art NR-IQA metrics. Moreover, the experimental results demonstrate that MDQM metric not only leads to high efficacy for classification of the various degradations but also maintains a remarkable run-time efficiency. It is worth to mention that the proposed method has been conducted for Arabic documents.

A Wavelet-Predominant Algorithm Can Evaluate Quality of THz Security Image and Identify Its Usability

This paper presents an aggregate wavelet-predominant algorithm to measure the distortions in THz security images. The algorithm integrates a spectral-based sharpness estimator, a noise estimator derived alpha-stable model and an overall viewing experience estimator based on free-energy principle. Among them, the greater weight is assigned to the spectral-based sharpness estimator considering that the main quality factor in THz security image is sharpness. To verify the feasibility of the proposed metric, we construct the THz security image dataset including a total of 181 THz security images, and each image has the mean opinion score (MOS) collected via subjective quality evaluation experiment. Quantitative experimental results on the constructed THz security image dataset show that the aggregate wavelet-predominant estimator produces the promising overall performance for the estimation of MOS values, with PLCC, SROCC, and RMSE of 0.900, 0.873, and 0.386, respectively. This performance is superior to other opinion-unaware approaches, viz. , FISBLIM, SISBLIM, NIQE, CPBD, SINE, S3, FISH, and noise estimator. The determination coefficient ( ${R} ^{{2}}$ ) of linear regression between reference and predicted MOSs is 0.81. The result of Bland–Altman analysis further validates that the aggregate wavelet-predominant estimator can substitute for the subjective IQA of THz security image, with approximately 94.5% of data points locating within the limits of agreement. For usability identification, the wavelet-predominant estimator gives the satisfactory results, with accuracy, precision, recall rate, and false positive rate of 84.0%, 79.8%, 95.0%, and 29.6%, respectively. Furthermore, the potential application perspectives of the proposed metric can refer to commercial applications (guarantee THz security images of good quality) and scientific researches (assist in software development for THz security image analysis). The dataset is available at https://doi.org/10.6084/m9.figshare.7700123.v3 . Possible researches on this dataset may include the development of THz quality standards, the selection of the best display mode, the enhancement of images, the modeling of image noise, and the detection of prohibited goods.

Quality assessment tool for performance measurement of image contrast enhancement methods

An objective image quality assessment tool is proposed to measure image enhancement quality with emphasis on contrast. In the proposed tool, which is based on maximizing contrast with minimum artefact (MCMA), local and global properties of an image are measured through pixel-wise and histogram-wise features, respectively. To this aim, three sub-measures are introduced, each of which able to detect one contrast-related quality aspect: (i) low dynamic range of image; (ii) histogram shape preservation during image enhancement process; and (iii) local pixel diversity. These sub-measures are combined through a subjective test to adapt them to the mean opinion scores (MOSs) of a diverse set of training contrast-enhanced images. A regression algorithm performs the adaptation by fitting the three sub-measures to the MOS values and finding an optimal linear combination by maximizing the Pearson correlation. In order to evaluate the performance of the MCMA algorithm, another independent, subsequent, subjective test was performed on a set of images enhanced by various known contrast enhancement algorithms to obtain MOS values and to compare them with the output of the proposed MCMA method. The experimental results show that MCMA has the highest correlation to the MOS when compared to the existing tested contrast measurement tools.

Adaptive QoE-based architecture on cloud mobile media for live streaming

Nowadays, more than 75% of Internet traffic is multimedia traffic, moreover mobile traffic is growing at a rate of 50% each year. All these data together with the evolution of the cloud infrastructures lead us to develop Cloud Mobile Media (CMM) architectures to support the needs demanded by end users. Nevertheless, due to an inherit higher and variable end to end delay mainly as a result of the virtualization process, new challenges appear in particular for live video streaming applications in order to keep a good Quality of Experience (QoE) of the delivered video. Thus, to keep client’s satisfaction within good levels in terms of Mean Opinion Score (MOS), we propose an adaptive QoE-based architecture running on CMM infrastructures for live streaming services. In order to carry out this goal, we propose an estimation of MOS values using an statistical method based on factor analysis. This estimation is based on different measured variables throughout the CMM infrastructure. In addition, we compare the accuracy of the estimated MOS against well-known publicly available video quality algorithms. With these estimations, our proposal is based on two added controllers to the CMM infrastructure: (a) the Software Defined Network controller that acts as a master and (b) the Media Streamer controller. Each one does different actions on the CMM infrastructure in order to maintain and improve the QoE at each end user. Finally, this architecture has been implemented over a fat tree topology in order to show their functionality. The results show that our proposal works properly and it adapts quickly to the network changes in order to deliver a good MOS.

Enhancing Image Quality via Style Transfer for Single Image Super-Resolution

Recently, by feat of the Generative Adversarial Network (GAN), single image super-resolution (SISR) has achieved great breakthroughs in enhancing the perceptual image quality. However, since the network is trained by minimizing the perceptual loss, the GAN based SISR method (SRGAN) [1] results in images with very low objective quality, i.e., peak signal-to-noise ratio (PSNR). In this letter, we aim to solve this problem in an image style transfer way, to generate an image with similar perceptual quality as SRGAN, but with much higher objective quality. Moreover, we propose a threshold-based method to automatically alter the objective and perceptual quality of the reconstructed image through adjusting only one parameter. Experimental results show that our method can achieve more than 1.6 dB PSNR improvement over SRGAN with similar Mean Opinion Score value. Also, with the same objective quality, our method can provide significantly better perceptual results than other state-of-the-art SISR methods.

Predicting VoLTE Quality using Random Neural Network

Long Term Evolution (LTE) was initially designed for a high data rates network. However, voice service is always a main service that drives huge profits benefit for mobile phone operators. Hence the deployment of Voice over LTE (VoLTE) is very essential. LTE network is a fully All-IP network, thus, the deployment of VoLTE is quite complex, specially for guaranteeing of Quality of Service (QoS) for meeting quality of experience of mobile users. The key purpose of this paper is to present an object, non-intrusive prediction model for VoLTE quality based on Random Neural Network (RNN). In order to simulate an experiment, a three-layer feed-forward RNN architecture with gradient descent training algorithm is applied. The inputs of this model are object network impairments such as Packet Loss Rate (PLR), Delay and Jitter. The VoLTE quality was predicted in term of the Mean Opinion Score (MOS). The simulation results show that this model offers MOS values which are quite close to well-known method is WB-PESQ (Wideband Perceptual Evaluation of Speech Quality) model. The results also show that the proposed model is very suitable for predicting voice quality over LTE network.