WO2021156525A1 - Method and system for estimating perceived quality in an audiovisual signal - Google Patents

Method and system for estimating perceived quality in an audiovisual signal Download PDF

Info

Publication number
WO2021156525A1
WO2021156525A1 PCT/ES2020/070809 ES2020070809W WO2021156525A1 WO 2021156525 A1 WO2021156525 A1 WO 2021156525A1 ES 2020070809 W ES2020070809 W ES 2020070809W WO 2021156525 A1 WO2021156525 A1 WO 2021156525A1
Authority
WO
WIPO (PCT)
Prior art keywords
audiovisual
quality
signal
computer
network
Prior art date
Application number
PCT/ES2020/070809
Other languages
Spanish (es)
French (fr)
Inventor
David JIMÉNEZ BERMEJO
Álvaro LLORENTE GÓMEZ
David MARTÍN GUTIÉRREZ
Andrés ARTALOYTIA VILARIÑO
José Manuel MENÉNDEZ GARCÍA
Federico ÁLVAREZ GARCÍA
Original Assignee
Universidad Politécnica de Madrid
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Universidad Politécnica de Madrid filed Critical Universidad Politécnica de Madrid
Publication of WO2021156525A1 publication Critical patent/WO2021156525A1/en

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/20—Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
    • H04N21/23—Processing of content or additional data; Elementary server operations; Server middleware
    • H04N21/234—Processing of video elementary streams, e.g. splicing of video streams or manipulating encoded video stream scene graphs
    • H04N21/23418—Processing of video elementary streams, e.g. splicing of video streams or manipulating encoded video stream scene graphs involving operations for analysing video streams, e.g. detecting features or characteristics
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/40—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks using virtualisation of network functions or resources, e.g. SDN or NFV entities
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L43/00—Arrangements for monitoring or testing data switching networks
    • H04L43/04—Processing captured monitoring data, e.g. for logfile generation
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L43/00—Arrangements for monitoring or testing data switching networks
    • H04L43/08—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L43/00—Arrangements for monitoring or testing data switching networks
    • H04L43/20—Arrangements for monitoring or testing data switching networks the monitoring system or the monitored elements being virtualised, abstracted or software-defined entities, e.g. SDN or NFV
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/20—Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
    • H04N21/23—Processing of content or additional data; Elementary server operations; Server middleware
    • H04N21/24—Monitoring of processes or resources, e.g. monitoring of server load, available bandwidth, upstream requests
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
    • H04N21/43—Processing of content or additional data, e.g. demultiplexing additional data from a digital video stream; Elementary client operations, e.g. monitoring of home network or synchronising decoder's clock; Client middleware
    • H04N21/44—Processing of video elementary streams, e.g. splicing a video clip retrieved from local storage with an incoming video stream or rendering scenes according to encoded video stream scene graphs
    • H04N21/44008—Processing of video elementary streams, e.g. splicing a video clip retrieved from local storage with an incoming video stream or rendering scenes according to encoded video stream scene graphs involving operations for analysing video streams, e.g. detecting features or characteristics in the video stream
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
    • H04N21/43—Processing of content or additional data, e.g. demultiplexing additional data from a digital video stream; Elementary client operations, e.g. monitoring of home network or synchronising decoder's clock; Client middleware
    • H04N21/442—Monitoring of processes or resources, e.g. detecting the failure of a recording device, monitoring the downstream bandwidth, the number of times a movie has been viewed, the storage space available from the internal hard disk
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/60—Network structure or processes for video distribution between server and client or between remote clients; Control signalling between clients, server and network components; Transmission of management data between server and client, e.g. sending from server to client commands for recording incoming content stream; Communication details between server and client 
    • H04N21/63—Control signaling related to video distribution between client, server and network components; Network processes for video distribution between server and clients or between remote clients, e.g. transmitting basic layer and enhancement layers over different transmission paths, setting up a peer-to-peer communication via Internet between remote STB's; Communication protocols; Addressing
    • H04N21/647—Control signaling between network components and server or clients; Network processes for video distribution between server and clients, e.g. controlling the quality of the video stream, by dropping packets, protecting content from unauthorised alteration within the network, monitoring of network load, bridging between two different networks, e.g. between IP and wireless
    • H04N21/64723—Monitoring of network processes or resources, e.g. monitoring of network load
    • H04N21/64738—Monitoring network characteristics, e.g. bandwidth, congestion level

Definitions

  • the invention falls within the Audiovisual Technologies sector for estimating perceived quality by applying an objective model of subjective quality of audiovisual content (signals) in media services for different digital content platforms with distribution through different communication networks.
  • the invention can be implemented as a software product, virtualizable and instantiable in any communication network, which offers the results of the final quality assessment of the analyzed content, through a user interface, on a limited numerical scale.
  • the invention allows the monitoring of the quality of the contents that are transported, and also provides relevant information that allows the optimization of the audiovisual services deployed in communication networks.
  • IP Television IP Television
  • Internet Protocol Television content associated with social networks
  • streaming and on-demand services virtual reality services
  • video games are responsible for increasing traffic network and with expectations of continuing exponential growth.
  • QoE assessment plays a key role for networked multimedia applications and services. Especially, through the development of objective QoE metrics that correlate with perceived subjective measures. To do this, they rely on different schemes to try to model how users perceive and experience quality losses in content, such as degradations. East Analysis is usually focused on the two main elements of multimedia services and applications: video and audio.
  • QoE models are based on the quality assessment (QA) of audio (AQA, Audio Quality Assessment), video (VQA, Video Quality Assessment) or both, analyzing the different processes that they can cause deterioration of signal quality such as: acquisition, preprocessing, encoding, transmission, presentation and storage.
  • QA quality assessment
  • AQA Audio Quality Assessment
  • VQA Video Quality Assessment
  • subjective methods can be distinguished, in which users rate the quality they perceive in the content offered under certain conditions and through standardized schemes; and objectives, which apply mathematical algorithms to the information they can extract from network media services and applications to approximate the measure of perceived quality.
  • Different objective quality models are normally used for audio and video.
  • the most common models estimate the degree of quality, or quality degradation, due to coding, taking into account parameters such as bit rate, resolution, number of frames per second, sampling rate, number of channels, etc. .
  • the usual result delivered by these quality models is typically a list of scores on a MOS (Mean Opinion Score) scale, where each score represents quality for a time segment of content.
  • MOS Molean Opinion Score
  • Some of these models are standardized in the recommendation of the International Telecommunications Union (ITU) P.1201.
  • ITU International Telecommunications Union
  • different categories are defined for the objective systems, from low to high complexity: parametric, bitstream and hybrids.
  • the objective models are compared with subjective measures of the same contents on which the objective measures are applied. These are carried out following standardized procedures, included in international regulations and recommendations, such as ITU-R BT.500 (Methodology for the subjective assessment of the quality of television pictures), ITU-T P.910 (Subjective video quality assessment methods for multimedia applications) and ITU-T P.913 (Methods for the subjective assessment of video quality, audio quality and audiovisual quality of Internet video and distribution quality television in any environmentf).
  • ITU-R BT.500 Methodhodology for the subjective assessment of the quality of television pictures
  • ITU-T P.910 Subjective video quality assessment methods for multimedia applications
  • ITU-T P.913 Methodhods for the subjective assessment of video quality, audio quality and audiovisual quality of Internet video and distribution quality television in any environmentf.
  • the invention belongs to the technical field of quality analysis of audiovisual content in media services for different digital content platforms with distribution through different communication networks.
  • the invention can be implemented as a software product, virtualizable and instantiable in any communication network, which offers the results of the quality assessment of the analyzed content, through a user interface, on a limited numerical scale.
  • audiovisual content services can come from different content distribution, broadcast and / or transport platforms.
  • the present invention has the ability to work with constant, variable and adaptive bit rates for all types of audiovisual signal.
  • Another distinctive feature of the present invention is that it has the ability to offer measurements in real time, and in configurable blocks of time. Also, the ability to work with any type of audiovisual signal, regardless of its nature, origin and / or configuration. Furthermore, the means of analysis and signal processing do not need references to the original signal.
  • the invention can be carried out by means of several interconnected modules, which are described below:
  • Capture module developed in a high-level programming language, which includes the transport stream decapsulator (TS, from English Transport Stream) and other audiovisual containers such as AVI or Matroska, streaming in any of its modalities (for example, UDP (User Datagram Protocol), RTP (Real Time Protocol), RTSP (Real Time Streaming Protocol), HTTP (Hypertext Transfer Protocol), unicast and multicast, adaptive streaming based on MPEG DASH (Dynamic Adaptive Streaming over HTTP), HLS (HTTP Uve Streaming), etc.) for multi-resolution content; and decoding of any type of audiovisual signal (that makes use, for example, of MPEG-2, MPEG-4 Part 10 (A VC or H.264), MPEG-H Part 2 (HEVC or H.265) encoding, etc.
  • TS transport stream decapsulator
  • AVI Real Time Streaming Protocol
  • HTTP Hypertext Transfer Protocol
  • MPEG DASH Dynamic Adaptive Streaming over HTTP
  • HLS HTTP Uve
  • Module of calculation of quality metrics without reference developed in a high-level programming language for the calculation of the perceptible levels of distortions associated with the degradation of the audiovisual signal due to the stages of acquisition, preprocessing, encoding, transmission , presentation and storage of s signal, and those introduced by the distribution network, broadcasting and / or transport of audiovisual content, such as: spatial complexity, temporal complexity, absence of audio signal, level of blurring in the video signal, speech intelligibility, level presence of blocks in the image, presence of frozen frames, loss of audio channels, excessive presence of black frames in the video, distortion level of high-frequency areas in the image, equal
  • These measurements are made on the audiovisual information obtained, and are offered as a set of significant measurements together with their relevant statistics that are offered as input for the next module in the form of a feature vector.
  • the range of potential values of the metrics is known, and allows setting thresholds to grant different treatments and priorities when they are transferred to the end user through the graphical interface (for example, through alarms), or through information compilation files. and generated by the invention, either created in real time, or through the management of structured data in the storage system of the system.
  • Quality prediction module based on mathematical algorithms, signal processing and machine learning techniques, which allows, on a programming model based on the implementation of software pipes (pipelines), the establishment of communications between the different sub-modules employees to attend to the user settings on the system.
  • the prediction module is built on the basis of mathematical algorithms, signal processing and machine learning techniques, the latter supported by training processes that use a wide set of labeled audiovisual content, which allow characterizing the presence of different artifacts, and the response that originate at the level of perceived or experienced quality at the time of consumption.
  • the sub-modules that the quality prediction module integrates are: Lasso and Ridge regressors, regressors with support machines (SVR), Random Forest and various deep learning models (Deep Learning).
  • the characteristics obtained from both the capture module (vector with the parametric properties of the signal) and the module of quality metrics without reference (vector with the values of the measurements per analyzed time segment and their statistics of up to second order) are used as input parameters for the quality prediction system, which, after passing through the different processing stages internal to the quality prediction module, end up offering, in the output layer, the quality result.
  • the prediction architectures used are oriented towards efficient computation to provide the system with the capacity to work in real time, and optimized in terms of precision and absence of errors.
  • One aspect of the present invention is a method for estimating perceived quality in audiovisual signal.
  • the audiovisual signal is transmitted by at least one communication network with audiovisual content selected from the distribution network, the broadcast network and the transport network. The method comprises the following steps:
  • the characteristic parameters of the service are selected between the resolution of the image, the number of channels and the sampling frequency of the audio, the bit rates used in the coding audio and video, and combinations of the above.
  • calculating a set of metrics without reference additionally comprises calculating an energy distribution as a function of frequency and detecting structures in both the audio signal and the signal. video with the ability to produce perceptual discomfort from encoding.
  • the mathematical algorithm, the signal processing and the machine learning techniques are Lasso and Ridge regressors, support machine regressors (SVR), Random Forest, and unvarious Deep Learning models.
  • FIG. 1 represents the functional blocks of the system. The drawing is made up of the following elements: 1) Capture module [1]
  • Figure 2 shows a flow chart for the three main modules:
  • Capture module responsible for the acquisition of the audiovisual signal, its most significant parameters and the parameters of the service [1]
  • Figure 3 shows a flow chart of data exchange at the operational level in the following stages.
  • the data exchanges are carried out through a data file with predefined coding.
  • Capture module [1] Formed by two modules that acquired the audiovisual flow, they separate the service metadata [11] and the media content [12] From them, the service information [13] and the information are extracted of audiovisual content [14] The first is intended for the quality prediction module [3], the second for the module of quality metrics without reference [2]
  • Non-referenced quality metrics module [2] Provides your results to the quality prediction module [3]
  • Quality prediction module [3] The quality result is transferred by means of a data file with predefined coding to the graphical user interface [5] and to the structured information storage system of the invention [4]
  • FIG. 4 shows a flow chart according to the present invention at the level of presentation of results. They are presented:
  • the method provides a value of the equivalent of perceived quality expressed on a limited numerical scale after a sequential process of applying a model that includes three stages aimed at: 1) capturing the information from the audio and video stream [1], 2) calculating a set of quality metrics without reference on audiovisual content [2], and 3) provide the measure of quality offered using this data in a model based on mathematical algorithms (which makes use of signal processing and machine learning) that provides a prediction of perceived quality according to various operating modes (predefined a priori and characterized by the parameters considered to make the prediction) [3]
  • the data flow object of analysis coming from a network of distribution, diffusion and / or transport of audiovisual content, is captured by means of a module that allows the collection of parameters of characterization of the media [14] and of the quality of the transport service of information [13]
  • the parameters captured are, among others, the resolution of the image, the number of frames per second, the number of channels and the sampling frequency of the audio, and the bit rates used in the encoding of audio and video.
  • the module captures the audiovisual signal for further processing.
  • This processing is carried out in the second [2] and third module [3], destined to obtain the parameters of the audiovisual signal, such as the video luminance levels and their distribution, the tint and saturation values, the component frequencies and audio formants, and entropy levels (both audio and video) and the calculation of quality metrics without reference [21] on it, focused on the calculation of the energy distribution as a function of frequency and the detection of structures both in the audio and in the image with the capacity to produce perceptual discomfort from the encoding.
  • the number of parameters considered varies depending on the previously selected operating mode, which allows selecting the precision (complexity) of the algorithm and, consequently, the speed of execution.
  • the parameterization obtained from the service [13], together with the parameters of the audiovisual signal [14] and the measurements of the quality metrics without reference [21], are inserted in a data file with predefined coding for its exchange between the modules of the system. All this information is turned against the quality prediction module [3] which, taking the input parameters compiled in the data file with predefined communication coding, returns a quality value on a limited numerical scale after the calculations performed in the prediction module, depending on the selected operating mode. To do this, it passes the data introduced by a set of mathematical models such as Lasso and Ridge regressors, regressors with support machines (SVR), Random Forest, and various Deep Learning models.
  • the result is dependent on the parameters obtained by means of mathematical algorithms, signal processing and applied machine learning techniques.
  • the system returns a numerical value on a limited scale, resulting from the processes executed by the different modules of the invention. Said numerical value constitutes an assessment of the quality of the audiovisual content under analysis per predetermined time unit.
  • the invention offers textual and graphic information temporally synchronized with the captured audiovisual signal, which allows the subsequent generation of quality reports and the generation of previously configurable alarms in the system [55]
  • the invention includes a structured storage system, which provides scalability, performance and high availability, for the entire set of analyzed audiovisual segments and the volume of data obtained over time [4]
  • the invention includes a graphical interface for visualization and interaction with the user [5] that allows the visualization of the evolution in real time of the quality measures provided by the system [54], the visualization of the temporal evolution of the same in a time interval set by the user [53], and access to a set of additional information, parameterizable by the user, with exact temporal correspondence with various effects potentially present in the stream of audiovisual content analyzed.
  • the QoE measurement represents a challenge not completely solved in the environment of the distribution, broadcasting and / or transport of audiovisual content, where the availability of the original content (to use as a reference) is very scarce or null, and the influence of those processed on the contents, required by the necessary adaptation to the capacities of the distribution, diffusion and / or transport network, very significant.
  • Subjective evaluations are unfeasible because they are expensive and slow, and current objective models lack the flexibility and scalability necessary to meet the growing needs related to the provision of media services and applications in order to ensure the best provision of them to users. end.
  • the invention proposes a virtualizable software service that can be dynamically instantiated at different points, end-to-end, of the multimedia content distribution, broadcast and / or transport network.
  • the invention makes it possible to provide a quality measure, similar to what a user could provide, making use exclusively of information from the network and the content itself, without the need to rely on versions of that content prior to the processing stages that precede it. in the chain of distribution, diffusion and / or transport of the same.
  • the quality measure offered by the system provided as a numerical value within a limited scale, comes from an algorithm developed with the help of mathematical algorithms, signal processing and machine learning techniques, and has the ability to self-adapt through the information it is able to collect during its operation, relying on reinforcement learning techniques.
  • the present invention overcomes the problems noted above through the features listed in the claims.
  • the invention allows, by means of a compact software, instantiable as a virtual machine or container, the performance of the calculation of the quality prediction of an audiovisual content with different properties (resolution in the case of video, encoding standard for both audio as for video, bits / audio sample, bit rate for both audio and video, etc.), offered as a numerical result on an easily intelligible scale, through a visual interface, which can also be stored and later retrieved in a structured way.
  • this proposal has: 1) the capacity to function in a virtualized way, allowing it to be dynamically instantiated at different points and with different time durations through different virtualization technical solutions; 2) ability to work on different content in parallel, or on the same content at different points in the distribution, broadcast and / or transport network of audiovisual content; 3) ability to capture and process the data resulting from its execution to improve its own measurement models, applying machine learning techniques by reinforcement; 4) ability to offer results in adjustable scales and even user-defined, taking into account different sensitivities to discrimination capabilities against possible defects present in the content; 5) ability to establish different storage structures for data, work with them individually and in aggregate, and retrieve them according to different criteria set by the user.
  • the object of the present invention is not any claim on the structured information storage system, nor the data presentation and visualization technologies linked to the graphical interface.
  • the application possibilities are manifold, since the invention offers numerous advantages.
  • the system allows the evaluation of the quality of audiovisual content, offering a result that can be used for: • Provide a value on a quality scale that allows the qualification and discrimination of audiovisual content services regardless of their nature and distribution, broadcast or transport platform.
  • the invention allows the realization of any application for monitoring and analyzing the quality of audiovisual content and audiovisual content services.

Landscapes

  • Engineering & Computer Science (AREA)
  • Signal Processing (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Multimedia (AREA)
  • Environmental & Geological Engineering (AREA)
  • Data Mining & Analysis (AREA)
  • Computer Security & Cryptography (AREA)
  • Databases & Information Systems (AREA)
  • Testing, Inspecting, Measuring Of Stereoscopic Televisions And Televisions (AREA)

Abstract

The present invention enables the quality of the contents that are transported to be monitored, and it further provides relevant information that enables the audiovisual services deployed in communication networks to be optimised. The method for estimating perceived quality in an audiovisual signal comprises the following steps: capturing characteristic parameters of the data transport service regardless of the network, both for the video signal and the audio signal contained in the audiovisual signal; calculating a set of no-reference metrics, from the characteristic parameters of the transport service, which provide an ordered set of parameters with the quality distortion levels of the audiovisual content; and, obtaining a perceived quality value by applying a mathematical algorithm, signal processing and machine learning techniques on the characteristic parameters of the service and the set of no-reference metrics.

Description

MÉTODO Y SISTEMA PARA LA ESTIMACIÓN DE CALIDAD PERCIBIDA EN SEÑAL METHOD AND SYSTEM FOR THE ESTIMATION OF PERCEIVED QUALITY IN SIGNAL
AUDIOVISUAL AUDIOVISUAL
DESCRIPCIÓN DESCRIPTION
Sector de la técnica Technical sector
La invención se encuadra en el sector de Tecnologías Audiovisuales para la estimación de la calidad percibida mediante la aplicación de un modelo objetivo de calidad subjetiva de contenidos (señales) audiovisuales en servicios de media para distintas plataformas digitales de contenidos con distribución mediante diferentes redes de comunicaciones. La invención se puede implementar como un producto software, virtualizable e instanciable en cualquier red de comunicaciones, que ofrece los resultados de la valoración de calidad final del contenido analizado, a través de una interfaz de usuario, en una escala numérica acotada. The invention falls within the Audiovisual Technologies sector for estimating perceived quality by applying an objective model of subjective quality of audiovisual content (signals) in media services for different digital content platforms with distribution through different communication networks. . The invention can be implemented as a software product, virtualizable and instantiable in any communication network, which offers the results of the final quality assessment of the analyzed content, through a user interface, on a limited numerical scale.
La invención permite la monitorización de la calidad de los contenidos que se transportan, y además facilita información relevante que permite la optimización de los servicios audiovisuales desplegados en redes de comunicaciones. The invention allows the monitoring of the quality of the contents that are transported, and also provides relevant information that allows the optimization of the audiovisual services deployed in communication networks.
Estado de la técnica State of the art
El consumo de contenidos audiovisuales a través de diferentes redes y plataformas digitales resulta en la actividad que genera mayor tráfico de datos en red y diversos modelos de negocio en los que proporcionar la mejor calidad de la experiencia (QoE, del inglés Quality of Experiencé) resulta clave. Los servicios multimedia en red, como los de Televisión IP (IPTV, Internet Protocol Televisión), los contenidos asociados a redes sociales, los servicios de streaming y bajo demanda, los servicios de realidad virtual, los videojuegos, son responsables de cada vez mayor tráfico de red y con expectativas de continuar con un crecimiento exponencial. The consumption of audiovisual content through different networks and digital platforms results in the activity that generates the highest network data traffic and various business models in which providing the best quality of experience (QoE) results key. Networked multimedia services, such as IP Television (IPTV, Internet Protocol Television), content associated with social networks, streaming and on-demand services, virtual reality services, video games, are responsible for increasing traffic network and with expectations of continuing exponential growth.
En este contexto, la evaluación de QoE ocupa un papel clave para los servicios y aplicaciones multimedia en red. Especialmente, a través del desarrollo de métricas objetivas de QoE que se correlacionen con medidas subjetivas percibidas. Para ello, se apoyan en diferentes esquemas para tratar de modelar cómo los usuarios perciben y experimentan las pérdidas de calidad en los contenidos, como las degradaciones. Este análisis suele estar centrado en los dos elementos principales de los servicios y aplicaciones multimedia: el vídeo y el audio. In this context, QoE assessment plays a key role for networked multimedia applications and services. Especially, through the development of objective QoE metrics that correlate with perceived subjective measures. To do this, they rely on different schemes to try to model how users perceive and experience quality losses in content, such as degradations. East Analysis is usually focused on the two main elements of multimedia services and applications: video and audio.
Una estimación precisa de QoE permite a los proveedores de los servicios y aplicaciones multimedia mejorar la provisión de los servicios, y optimizar su uso de la red. De hecho, la evaluación de QoE de los contenidos audiovisuales se está convirtiendo en un tema de extrema importancia. An accurate estimate of QoE allows providers of multimedia services and applications to improve the provision of services, and optimize their use of the network. In fact, the QoE evaluation of audiovisual content is becoming an extremely important issue.
La mayoría de los modelos existentes de QoE se basan en la evaluación de calidad (QA, del inglés Quality Assessment) de audio (AQA, Audio Quality Assessment), video (VQA, Video Quality Assessment) o de ambos, analizando los diferentes procesos que pueden producir un deterioro de la calidad de la señal como: la adquisición, el preprocesamiento, la codificación, la transmisión, la presentación y el almacenamiento. Dentro de estos se pueden distinguir los métodos subjetivos, en los que usuarios califican la calidad que perciben en los contenidos ofrecidos bajo unas condiciones y a través de unos esquemas estandarizados; y los objetivos, que aplican algoritmias matemáticas a la información que pueden extraer de los servicios y aplicaciones de media en red para aproximar la medida de la calidad percibida. Most of the existing QoE models are based on the quality assessment (QA) of audio (AQA, Audio Quality Assessment), video (VQA, Video Quality Assessment) or both, analyzing the different processes that they can cause deterioration of signal quality such as: acquisition, preprocessing, encoding, transmission, presentation and storage. Within these, subjective methods can be distinguished, in which users rate the quality they perceive in the content offered under certain conditions and through standardized schemes; and objectives, which apply mathematical algorithms to the information they can extract from network media services and applications to approximate the measure of perceived quality.
Normalmente se utilizan diferentes modelos de calidad objetiva para audio y video. Los modelos más comunes estiman el grado de calidad, o el de degradación de calidad, debido a la codificación atendiendo a parámetros como la velocidad binaria, la resolución, el número de cuadros por segundo, la tasa de muestreo, el número de canales, etc. El resultado habitual entregado por estos modelos de calidad es típicamente una lista de puntuaciones en una escala MOS (del inglés Mean Opinión Score), donde cada puntuación representa la calidad para un segmento temporal de contenidos. Algunos de estos modelos se encuentran normalizados en la recomendación de la Unión Internacional de Telecomunicaciones (UIT) P.1201. Dependiendo de la tipología de los parámetros empleados, se definen diferentes categorías para los sistemas objetivos, de menor a mayor complejidad: paramétricos, de flujo de bits e híbridos. Different objective quality models are normally used for audio and video. The most common models estimate the degree of quality, or quality degradation, due to coding, taking into account parameters such as bit rate, resolution, number of frames per second, sampling rate, number of channels, etc. . The usual result delivered by these quality models is typically a list of scores on a MOS (Mean Opinion Score) scale, where each score represents quality for a time segment of content. Some of these models are standardized in the recommendation of the International Telecommunications Union (ITU) P.1201. Depending on the typology of the parameters used, different categories are defined for the objective systems, from low to high complexity: parametric, bitstream and hybrids.
Los modelos objetivos se comparan con medidas subjetivas de los mismos contenidos sobre los que se aplican las medidas objetivas. Éstas se realizan siguiendo procedimientos normalizados, recogidos en normativas y recomendaciones internacionales, como UIT-R BT.500 ( Methodology for the subjective assessment of the quality of televisión pictures), UIT-T P.910 ( Subjective video quality assessment methods for multimedia applications) y UIT-T P.913 ( Methods for the subjective assessment of video quality, audio quality and audiovisual quality of Internet video and distribution quality televisión in any environmenf). The objective models are compared with subjective measures of the same contents on which the objective measures are applied. These are carried out following standardized procedures, included in international regulations and recommendations, such as ITU-R BT.500 (Methodology for the subjective assessment of the quality of television pictures), ITU-T P.910 (Subjective video quality assessment methods for multimedia applications) and ITU-T P.913 (Methods for the subjective assessment of video quality, audio quality and audiovisual quality of Internet video and distribution quality television in any environmentf).
En cuanto a las patentes relacionadas con la calidad de la experiencia de las señales audiovisuales, cabe destacar la solicitud de patente CN108900862-A con título: “A network video stream QoE-QoS parameter mapping method based on statistics anaiysis”, y la solicitud de patente con número de publicación US2019124375-A1 con título: “Quality Estimation Of Adaptive Multimedia Streaming”. Regarding the patents related to the quality of the experience of audiovisual signals, it is worth highlighting the patent application CN108900862-A with title: “A network video stream QoE-QoS parameter mapping method based on statistics analysis”, and the application for Patent with publication number US2019124375-A1 with title: “Quality Estimation Of Adaptive Multimedia Streaming”.
Breve descripción de la invención Brief description of the invention
La invención pertenece al campo técnico del análisis de calidad de contenidos audiovisuales en servicios de media para diferentes plataformas digitales de contenidos con distribución mediante diferentes redes de comunicaciones. La invención se puede implementar como un producto software, virtualizable e instanciable en cualquier red de comunicaciones, que ofrece los resultados de la valoración de calidad del contenido analizado, a través de una interfaz de usuario, en una escala numérica acotada. The invention belongs to the technical field of quality analysis of audiovisual content in media services for different digital content platforms with distribution through different communication networks. The invention can be implemented as a software product, virtualizable and instantiable in any communication network, which offers the results of the quality assessment of the analyzed content, through a user interface, on a limited numerical scale.
En el entorno de la presente invención, los servicios de contenidos audiovisuales pueden proceder de diferentes plataformas de distribución, difusión y/o transporte de contenidos. Además, la presente invención tiene la capacidad de trabajar con tasas binarias constantes, variables y adaptativas para todo tipo de señal audiovisual. In the context of the present invention, audiovisual content services can come from different content distribution, broadcast and / or transport platforms. Furthermore, the present invention has the ability to work with constant, variable and adaptive bit rates for all types of audiovisual signal.
Otra característica distintiva de la presente invención es que tiene la capacidad de ofrecer las medidas en tiempo real, y en bloques configurables de tiempo. También, la capacidad de trabajar con cualquier tipo de señal audiovisual, independientemente de su naturaleza, procedencia y/o configuración. Además, los medios de análisis y procesado de la señal no necesitan de referencias de la señal original. Another distinctive feature of the present invention is that it has the ability to offer measurements in real time, and in configurable blocks of time. Also, the ability to work with any type of audiovisual signal, regardless of its nature, origin and / or configuration. Furthermore, the means of analysis and signal processing do not need references to the original signal.
La invención se puede llevar a cabo mediante varios módulos interconectados entre sí, los cuales se describen a continuación: The invention can be carried out by means of several interconnected modules, which are described below:
1) Módulo de captura: desarrollado en un lenguaje de programación de alto nivel, que incluye el desencapsulador para flujo de transporte (TS, del inglés Transport Stream) y otros contenedores audiovisuales como AVI o Matroska, streaming en cualquiera de sus modalidades (por ejemplo, UDP (User Datagram Protocoí), RTP (Real Time Protocoí), RTSP (Real Time Streaming Protocoí), HTTP (Hypertext Transfer Protocoí), unicast y multicast, streaming adaptativo basado en MPEG DASH (Dynamic Adaptive Streaming over HTTP), HLS (HTTP Uve Streaming), etc.) para contenidos multiresolución; y decodificación de cualquier tipo de señal audiovisual (que haga uso, por ejemplo, de codificación MPEG-2, MPEG-4 Parte 10 (A VC ó H.264), MPEG-H Parte 2 (HEVC ó H.265), etc. para vídeo, y MP3, AC3, AC3+, AC4, etc. para audio. Permite la extracción de parámetros que caracterizan el servicio de contenidos audiovisuales, como la resolución de la señal de vídeo, el estándar de codificación para audio y vídeo (por separado) empleado, la tasa binaria empleada para vídeo y audio, el tipo de escaneo y el número de cuadros por segundo en el caso del vídeo, el número de canales de audio, la frecuencia de muestreo empleada para el audio, etc. ) Módulo de cálculo de métricas de calidad sin referencia: desarrollado en un lenguaje de programación de alto nivel para el cálculo de los niveles perceptibles de distorsiones asociadas a la degradación de la señal audiovisual debida a las etapas de adquisición, el preprocesamiento, la codificación, la transmisión, la presentación y el almacenamiento de señal, y aquellos introducidos por la red de distribución, difusión y/o transporte de contenidos audiovisuales, tales como: complejidad espacial, complejidad temporal, ausencia de señal de audio, nivel de emborronado en la señal de vídeo, inteligibilidad de la voz, nivel de presencia de bloques en la imagen, presencia de cuadros congelados, pérdida de canales de audio, presencia excesiva de cuadros negros en el vídeo, nivel de distorsión de zonas de alta frecuencia en la imagen, desajustes de ecualización en los canales de audio, existencia de niveles col orí métricos desajustados en el vídeo, etc.. Estas medidas se realizan sobre la información audiovisual obtenida, y se ofrecen como un conjunto de medidas significativas junto con sus estadísticos relevantes que se ofrecen como entrada para el siguiente módulo en forma de un vector de características. El rango de valores potenciales de las métricas es conocido, y permite fijar umbrales para otorgar distintos tratamientos y prioridades a la hora de ser trasladados al usuario final a través de la interfaz gráfica (por ejemplo, mediante alarmas), o mediante archivos recopilatorios de información y generados por la invención, bien creados en tiempo real, bien a través de la gestión de los datos estructurados en el sistema de almacenamiento del sistema. 1) Capture module: developed in a high-level programming language, which includes the transport stream decapsulator (TS, from English Transport Stream) and other audiovisual containers such as AVI or Matroska, streaming in any of its modalities (for example, UDP (User Datagram Protocol), RTP (Real Time Protocol), RTSP (Real Time Streaming Protocol), HTTP (Hypertext Transfer Protocol), unicast and multicast, adaptive streaming based on MPEG DASH (Dynamic Adaptive Streaming over HTTP), HLS (HTTP Uve Streaming), etc.) for multi-resolution content; and decoding of any type of audiovisual signal (that makes use, for example, of MPEG-2, MPEG-4 Part 10 (A VC or H.264), MPEG-H Part 2 (HEVC or H.265) encoding, etc. . for video, and MP3, AC3, AC3 +, AC4, etc. for audio. It allows the extraction of parameters that characterize the audiovisual content service, such as the resolution of the video signal, the coding standard for audio and video (for separately) used, the bit rate used for video and audio, the type of scan and the number of frames per second in the case of video, the number of audio channels, the sample rate used for audio, etc.) Module of calculation of quality metrics without reference: developed in a high-level programming language for the calculation of the perceptible levels of distortions associated with the degradation of the audiovisual signal due to the stages of acquisition, preprocessing, encoding, transmission , presentation and storage of s signal, and those introduced by the distribution network, broadcasting and / or transport of audiovisual content, such as: spatial complexity, temporal complexity, absence of audio signal, level of blurring in the video signal, speech intelligibility, level presence of blocks in the image, presence of frozen frames, loss of audio channels, excessive presence of black frames in the video, distortion level of high-frequency areas in the image, equalization imbalances in the audio channels, existence of mismatched col orimetric levels in the video, etc. These measurements are made on the audiovisual information obtained, and are offered as a set of significant measurements together with their relevant statistics that are offered as input for the next module in the form of a feature vector. The range of potential values of the metrics is known, and allows setting thresholds to grant different treatments and priorities when they are transferred to the end user through the graphical interface (for example, through alarms), or through information compilation files. and generated by the invention, either created in real time, or through the management of structured data in the storage system of the system.
3) Módulo de predicción de calidad: basado en algoritmia matemática, procesado de señal y técnicas de aprendizaje máquina, que permite, sobre un modelo de programación basado en la implementación de tuberías software (pipelines), el establecimiento de las comunicaciones entre los diferentes submódulos empleados para atender a las parametrizaciones de usuario sobre el sistema. El módulo de predicción se construye en base a algoritmia matemática, procesado de señal y técnicas de aprendizaje máquina, estas últimas apoyadas en procesos de entrenamiento que emplean un conjunto amplio de contenidos audiovisuales, etiquetados, que permiten caracterizar la presencia de diferentes artefactos, y la respuesta que originan a nivel de calidad percibida o experimentada en el momento de su consumo. Los submódulos que integra el módulo de predicción de calidad son: regresores de Lasso y Ridge, regresores con máquinas de soporte (SVR), Random Forest y varios modelos de aprendizaje profundo ( Deep Learning ). 3) Quality prediction module: based on mathematical algorithms, signal processing and machine learning techniques, which allows, on a programming model based on the implementation of software pipes (pipelines), the establishment of communications between the different sub-modules employees to attend to the user settings on the system. The prediction module is built on the basis of mathematical algorithms, signal processing and machine learning techniques, the latter supported by training processes that use a wide set of labeled audiovisual content, which allow characterizing the presence of different artifacts, and the response that originate at the level of perceived or experienced quality at the time of consumption. The sub-modules that the quality prediction module integrates are: Lasso and Ridge regressors, regressors with support machines (SVR), Random Forest and various deep learning models (Deep Learning).
Las características obtenidas tanto del módulo de captura (vector con las propiedades paramétricas de la señal) como del módulo de métricas de calidad sin referencia (vector con los valores de las medidas por segmento temporal analizado y sus estadísticos de hasta segundo orden) se emplean como parámetros de entrada para el sistema de predicción de calidad, que, tras el paso por las diferentes etapas de procesado internas al módulo de predicción de calidad, terminan ofreciendo, en la capa de salida, el resultado de calidad. Las arquitecturas de predicción empleadas están orientadas a la computación eficiente para dotar al sistema de capacidad de trabajo en tiempo real, y optimizadas en términos de precisión y ausencia de errores. The characteristics obtained from both the capture module (vector with the parametric properties of the signal) and the module of quality metrics without reference (vector with the values of the measurements per analyzed time segment and their statistics of up to second order) are used as input parameters for the quality prediction system, which, after passing through the different processing stages internal to the quality prediction module, end up offering, in the output layer, the quality result. The prediction architectures used are oriented towards efficient computation to provide the system with the capacity to work in real time, and optimized in terms of precision and absence of errors.
Cuando la presente invención se implementa como un sistema embebido en un software compacto, instanciable como una máquina virtual o contenedor, permite la realización del cálculo de la predicción de calidad de un contenido audiovisual sin restricciones software o hardware impuestas por el entorno de trabajo donde se despliegue. Además, el despliegue de la presente invención se puede realizar en cualquier punto de la cadena de distribución, difusión y/o transporte, extremo a extremo. Un aspecto de la presente invención es un método para la estimación de calidad percibida en señal audiovisual. La señal audiovisual se transmite por al menos una red de comunicaciones con contenidos audiovisuales seleccionada entre red de distribución, red de difusión y red de transporte. El método comprende los siguientes pasos: When the present invention is implemented as a system embedded in compact software, instantiable as a virtual machine or container, it allows the calculation of the quality prediction of an audiovisual content to be carried out without software or hardware restrictions imposed by the work environment where it is used. deployment. Furthermore, the deployment of the present invention can be carried out at any point in the distribution, diffusion and / or transport chain, end-to-end. One aspect of the present invention is a method for estimating perceived quality in audiovisual signal. The audiovisual signal is transmitted by at least one communication network with audiovisual content selected from the distribution network, the broadcast network and the transport network. The method comprises the following steps:
- capturar unos parámetros característicos del servicio de transporte de datos con independencia de la red, tanto para la señal de vídeo como la señal de audio contenidas en la señal audiovisual; - capture some characteristic parameters of the data transport service regardless of the network, both for the video signal and the audio signal contained in the audiovisual signal;
- calcular un conjunto de métricas sin referencia, a partir de los parámetros característicos del servicio de transporte de datos, que proporcionan un conjunto ordenado de parámetros con los niveles de distorsión de calidad de los contenidos audiovisuales; y, - calculate a set of metrics without reference, from the characteristic parameters of the data transport service, which provide an ordered set of parameters with the quality distortion levels of the audiovisual content; Y,
- obtener un valor de calidad percibida aplicando algoritmia matemática, procesado de señal y técnicas de aprendizaje máquina sobre los parámetros característicos del servicio y el conjunto de métricas sin referencia. - Obtain a perceived quality value by applying mathematical algorithms, signal processing and machine learning techniques on the characteristic parameters of the service and the set of metrics without reference.
En una forma de realización del método para la estimación de calidad percibida en señal audiovisual, los parámetros característicos del servicio están seleccionados entre la resolución de la imagen, el número de canales y la frecuencia de muestreo del audio, las tasas binarias empleadas en la codificación de audio y de vídeo, y combinaciones de los anteriores. In an embodiment of the method for estimating perceived quality in audiovisual signal, the characteristic parameters of the service are selected between the resolution of the image, the number of channels and the sampling frequency of the audio, the bit rates used in the coding audio and video, and combinations of the above.
En otra forma de realización del método para la estimación de calidad percibida en señal audiovisual, calcular un conjunto de métricas sin referencia adicionalmente comprende calcular una distribución de energía en función de la frecuencia y detectar unas estructuras tanto en la señal de audio como en la señal de video con capacidad de producir incomodidad perceptual procedentes de la codificación. In another embodiment of the method for estimating perceived quality in audiovisual signal, calculating a set of metrics without reference additionally comprises calculating an energy distribution as a function of frequency and detecting structures in both the audio signal and the signal. video with the ability to produce perceptual discomfort from encoding.
En otra forma de realización del método para la estimación de calidad percibida en señal audiovisual, la algoritmia matemática, el procesado de señal y las técnicas de aprendizaje máquina son regresores de Lasso y Ridge, regresores con máquinas de soporte (SVR), Random Forest, y unvarios modelos de Deep Learning. In another embodiment of the method for the estimation of perceived quality in audiovisual signal, the mathematical algorithm, the signal processing and the machine learning techniques are Lasso and Ridge regressors, support machine regressors (SVR), Random Forest, and unvarious Deep Learning models.
Breve descripción de las figuras Brief description of the figures
La Figura 1 representa los bloques funcionales del sistema. El dibujo está compuesto por los siguientes elementos: 1) Módulo de captura [1] Figure 1 represents the functional blocks of the system. The drawing is made up of the following elements: 1) Capture module [1]
2) Módulo de cálculo de métricas de calidad sin referencia [2] 2) Module for calculating quality metrics without reference [2]
3) Módulo de predicción de calidad [3] 3) Quality prediction module [3]
4) Intercambio de información entre módulos. 4) Exchange of information between modules.
5) Sistema estructurado de almacenamiento de información [4] 5) Structured information storage system [4]
6) Intercambio de información con el sistema estructurado de almacenamiento de información. 6) Exchange of information with the structured information storage system.
7) Interfaz de usuario [5] 7) User interface [5]
8) Intercambio de información con la interfaz. 8) Information exchange with the interface.
La Figura 2 muestra un diagrama de flujo para los tres módulos principales: Figure 2 shows a flow chart for the three main modules:
1) Módulo de captura. Responsable de la adquisición de la señal audiovisual, sus parámetros más significativos y los parámetros del servicio [1] 1) Capture module. Responsible for the acquisition of the audiovisual signal, its most significant parameters and the parameters of the service [1]
2) Módulo de métricas de calidad sin referencia. Proporciona medidas de calidad del contenido audiovisual capturado [2] 2) Module of quality metrics without reference. Provides quality measures of captured audiovisual content [2]
3) Módulo de predicción de calidad. Proporciona el resultado de calidad obtenido [3] 3) Quality prediction module. Provides the quality result obtained [3]
La Figura 3 muestra un diagrama de flujo de intercambio de datos a nivel operativo en las etapas siguientes. Los intercambios de datos se realizan a través de fichero de datos con codificación predefinida. Figure 3 shows a flow chart of data exchange at the operational level in the following stages. The data exchanges are carried out through a data file with predefined coding.
1) Módulo de captura [1] Formado por dos módulos que adquirido el flujo audiovisual, separan la metadata de servicio [11] y el contenido de media [12] A partir de ellos se extraen la información del servicio [13] y la información del contenido audiovisual [14] La primera se destina al módulo de predicción de la calidad [3], la segunda al módulo de métricas de calidad sin referencia [2] 1) Capture module [1] Formed by two modules that acquired the audiovisual flow, they separate the service metadata [11] and the media content [12] From them, the service information [13] and the information are extracted of audiovisual content [14] The first is intended for the quality prediction module [3], the second for the module of quality metrics without reference [2]
2) Módulo de métricas de calidad sin referencia [2] Proporciona sus resultados al módulo de predicción de la calidad [3] 2) Non-referenced quality metrics module [2] Provides your results to the quality prediction module [3]
3) Módulo de predicción de la calidad [3] El resultado de calidad se traslada mediante un fichero de datos con codificación predefinida a la interfaz gráfica de usuario [5] y al sistema de almacenamiento estructurado de información de la invención [4] 3) Quality prediction module [3] The quality result is transferred by means of a data file with predefined coding to the graphical user interface [5] and to the structured information storage system of the invention [4]
La Figura 4 muestra un diagrama de flujo de acuerdo con la presente invención a nivel de presentación resultados. Se presentan: Figure 4 shows a flow chart according to the present invention at the level of presentation of results. They are presented:
1) Información gráfica y textual de la fuente (contenido audiovisual) bajo análisis 1) Graphic and textual information of the source (audiovisual content) under analysis
[51]. 2) Acceso al sistema de almacenamiento estructurado del sistema [52] [51]. 2) Access to the system's structured storage system [52]
3) Información gráfica de la calidad del contenido audiovisual en tiempo real procedente del módulo de predicción de la calidad [53] 3) Graphic information on the quality of audiovisual content in real time from the quality prediction module [53]
4) Información gráfica del histórico de calidad del contenido audiovisual en una horquilla de tiempos [54] 4) Graphic information on the quality history of audiovisual content in a time bracket [54]
5) Información gráfica y textual de los artefactos más significativos sufridos por el contenido audiovisual en el periodo de análisis categorizados como alarmas y avisos [55] 5) Graphic and textual information of the most significant artifacts suffered by audiovisual content in the analysis period, categorized as alarms and warnings [55]
Descripción detallada de un modo de realización Detailed description of an embodiment
Se refiere específicamente a un método basado en el análisis híbrido de propiedades paramétricas, las que se pueden extraer de la parametrización que se emplea en el flujo audiovisual, e intrínsecas de media, que incluye la captura de señal multimedia [1], la extracción de parámetros objetivos de la señal audiovisual y el cálculo de métricas objetivas [2], para proporcionar a través de algoritmia matemática (que hace uso de proceso de señal y aprendizaje máquina), con los elementos anteriormente descritos como entrada, una medida de calidad audiovisual [3] de un flujo audiovisual en red para diferentes plataformas de distribución, difusión y/o transporte de contenidos audiovisuales que pueden funcionar con sistemas de ancho de banda constante o variable y tasas binarias constantes, variables y adaptativas. It specifically refers to a method based on the hybrid analysis of parametric properties, which can be extracted from the parameterization used in the audiovisual stream, and intrinsic media, which includes the capture of multimedia signal [1], the extraction of objective parameters of the audiovisual signal and the calculation of objective metrics [2], to provide through mathematical algorithm (which makes use of signal processing and machine learning), with the elements previously described as input, a measure of audiovisual quality [ 3] of an audiovisual network flow for different platforms for the distribution, diffusion and / or transport of audiovisual content that can work with systems of constant or variable bandwidth and constant, variable and adaptive binary rates.
El método proporciona un valor del equivalente de calidad percibida expresado en una escala numérica acotada tras un proceso secuencial de aplicación de un modelo que incluye tres etapas destinadas a: 1) capturar la información del flujo de audio y vídeo [1], 2) calcular un conjunto de métricas de calidad sin referencia sobre el contenido audiovisual [2], y 3) proporcionar la medida de calidad ofrecida empleando estos datos en un modelo basado en algoritmia matemática (que hace uso de proceso de señal y aprendizaje máquina) que proporciona una predicción de la calidad percibida según varios modos de funcionamiento (predefinidos a priori y caracterizados por los parámetros considerados para hacer la predicción) [3] The method provides a value of the equivalent of perceived quality expressed on a limited numerical scale after a sequential process of applying a model that includes three stages aimed at: 1) capturing the information from the audio and video stream [1], 2) calculating a set of quality metrics without reference on audiovisual content [2], and 3) provide the measure of quality offered using this data in a model based on mathematical algorithms (which makes use of signal processing and machine learning) that provides a prediction of perceived quality according to various operating modes (predefined a priori and characterized by the parameters considered to make the prediction) [3]
El flujo de datos objeto de análisis, procedente de una red de distribución, difusión y/o transporte de contenidos audiovisuales, se captura mediante un módulo que permite la recopilación de parámetros de caracterización del media [14] y de la calidad del servicio de transporte de información [13] Los parámetros capturados son, entre otros, la resolución de la imagen, el número de cuadros por segundo, el número de canales y la frecuencia de muestreo del audio, y las tasas binarias empleadas en la codificación de audio y de vídeo. Además de los parámetros anteriores, el módulo captura la señal audiovisual para su posterior procesamiento. Este procesado se realiza en el segundo [2] y tercer módulo [3], destinados a la obtención de los parámetros de la señal audiovisual, como los niveles de luminancia de vídeo y su distribución, los valores de tinte y saturación, los valores de las frecuencias componentes y los formantes de audio, y los niveles de entropía (tanto de audio como de vídeo) y al cálculo de las métricas de calidad sin referencia [21] sobre la misma, centradas en el cálculo de la distribución de energía en función de la frecuencia y la detección de estructuras tanto en el audio como en la imagen con capacidad de producir incomodidad perceptual procedentes de la codificación. El número de parámetros considerados varía en función del modo de funcionamiento seleccionado previamente, que permite seleccionar la precisión (complejidad) de la algoritmia y, consecuentemente, la velocidad de ejecución. Dichos modos de funcionamiento permiten la operación del sistema en tiempo real independientemente de los recursos hardware y software disponibles en el entorno de despliegue del sistema. La parametrización obtenida del servicio [13], junto con los parámetros de la señal audiovisual [14] y las medidas de las métricas de calidad sin referencia [21], se inserta en un fichero de datos con codificación predefinida para su intercambio entre los módulos del sistema. Toda esta información se vuelca contra el módulo de predicción de calidad [3] que, tomando los parámetros de entrada compilados en el fichero de datos con codificación predefinida de comunicación, devuelve un valor de calidad en una escala numérica acotada tras los cálculos realizados en el módulo de predicción, según el modo de funcionamiento seleccionado. Para ello, hace pasar los datos introducidos por un conjunto de modelos matemáticos como regresores de Lasso y Ridge, regresores con máquinas de soporte (SVR), Random Forest, y varios modelos de Deep Learning. Para cualquiera de los modos de funcionamiento, el resultado es dependiente de los parámetros obtenidos mediante algoritmia matemática, procesado de señal y técnicas de aprendizaje máquina aplicadas. El sistema devuelve un valor numérico en una escala acotada, resultante de los procesos ejecutados por los diferentes módulos de la invención. Dicho valor numérico constituye una valoración de la calidad del contenido audiovisual bajo análisis por unidad de tiempo prefijada. Adicionalmente, la invención ofrece información textual y gráfica sincronizada temporalmente con la señal audiovisual capturada, que permite la posterior generación de informes de calidad y la generación de alarmas, previamente configurables, en el sistema [55] La invención incluye un sistema de almacenamiento estructurado, que proporciona escalabilidad, rendimiento y gran disponibilidad, para todo el conjunto de segmentos audiovisuales analizados y el volumen de datos que se obtiene a lo largo del tiempo [4] The data flow object of analysis, coming from a network of distribution, diffusion and / or transport of audiovisual content, is captured by means of a module that allows the collection of parameters of characterization of the media [14] and of the quality of the transport service of information [13] The parameters captured are, among others, the resolution of the image, the number of frames per second, the number of channels and the sampling frequency of the audio, and the bit rates used in the encoding of audio and video. In addition to the above parameters, the module captures the audiovisual signal for further processing. This processing is carried out in the second [2] and third module [3], destined to obtain the parameters of the audiovisual signal, such as the video luminance levels and their distribution, the tint and saturation values, the component frequencies and audio formants, and entropy levels (both audio and video) and the calculation of quality metrics without reference [21] on it, focused on the calculation of the energy distribution as a function of frequency and the detection of structures both in the audio and in the image with the capacity to produce perceptual discomfort from the encoding. The number of parameters considered varies depending on the previously selected operating mode, which allows selecting the precision (complexity) of the algorithm and, consequently, the speed of execution. These modes of operation allow the operation of the system in real time regardless of the hardware and software resources available in the system's deployment environment. The parameterization obtained from the service [13], together with the parameters of the audiovisual signal [14] and the measurements of the quality metrics without reference [21], are inserted in a data file with predefined coding for its exchange between the modules of the system. All this information is turned against the quality prediction module [3] which, taking the input parameters compiled in the data file with predefined communication coding, returns a quality value on a limited numerical scale after the calculations performed in the prediction module, depending on the selected operating mode. To do this, it passes the data introduced by a set of mathematical models such as Lasso and Ridge regressors, regressors with support machines (SVR), Random Forest, and various Deep Learning models. For any of the operating modes, the result is dependent on the parameters obtained by means of mathematical algorithms, signal processing and applied machine learning techniques. The system returns a numerical value on a limited scale, resulting from the processes executed by the different modules of the invention. Said numerical value constitutes an assessment of the quality of the audiovisual content under analysis per predetermined time unit. Additionally, the invention offers textual and graphic information temporally synchronized with the captured audiovisual signal, which allows the subsequent generation of quality reports and the generation of previously configurable alarms in the system [55] The invention includes a structured storage system, which provides scalability, performance and high availability, for the entire set of analyzed audiovisual segments and the volume of data obtained over time [4]
La invención incluye una interfaz gráfica de visualización e interacción con el usuario [5] que permite la visualización de la evolución en tiempo real de las medidas de calidad proporcionadas por el sistema [54], la visualización de la evolución temporal de las mismas en un intervalo de tiempo fijado por el usuario [53], y el acceso a un conjunto de informaciones adicionales, parametrizables por el usuario, con correspondencia temporal exacta con diversos efectos potencialmente presentes en el flujo de contenidos audiovisuales analizado. The invention includes a graphical interface for visualization and interaction with the user [5] that allows the visualization of the evolution in real time of the quality measures provided by the system [54], the visualization of the temporal evolution of the same in a time interval set by the user [53], and access to a set of additional information, parameterizable by the user, with exact temporal correspondence with various effects potentially present in the stream of audiovisual content analyzed.
La medida de QoE supone un reto no completamente resuelto en el entorno de la de distribución, difusión y/o transporte de contenidos audiovisuales, donde la disponibilidad de los contenidos originales (para usar como referencia) es muy escasa o nula, y la influencia de los procesados sobre los contenidos, requeridos por la necesaria adaptación a las capacidades de la red de distribución, difusión y/o transporte, muy significativa. Las valoraciones subjetivas son inviables por costosas y lentas, y los modelos objetivos actuales carecen de la flexibilidad y escalabilidad necesaria para atender las crecientes necesidades ligadas a la provisión de servicios y aplicaciones de media de cara a asegurar la mejor prestación de los mismos a los usuarios finales. The QoE measurement represents a challenge not completely solved in the environment of the distribution, broadcasting and / or transport of audiovisual content, where the availability of the original content (to use as a reference) is very scarce or null, and the influence of those processed on the contents, required by the necessary adaptation to the capacities of the distribution, diffusion and / or transport network, very significant. Subjective evaluations are unfeasible because they are expensive and slow, and current objective models lack the flexibility and scalability necessary to meet the growing needs related to the provision of media services and applications in order to ensure the best provision of them to users. end.
Como respuesta al reto anterior, la invención plantea un servicio software virtualizable que puede instanciarse dinámicamente en distintos puntos, extremo a extremo, de la red de distribución, difusión y/o transporte de contenidos multimedia. La invención permite proporcionar una medida de calidad, similar a la que un usuario podría proporcionar, haciendo uso exclusivamente de información de la red y del contenido en sí mismo, sin necesidad de apoyarse en versiones de ese contenido anteriores a etapas de procesado que le precedan en la cadena de distribución, difusión y/o transporte del mismo. La medida de la calidad ofrecida por el sistema, proporcionada como un valor numérico dentro de una escala acotada, procede de un algoritmo desarrollado con la ayuda de algoritmia matemática, procesado de señal y técnicas de aprendizaje máquina, y posee capacidad de auto-adaptarse mediante la información que es capaz de recopilar durante su funcionamiento, apoyándose en técnicas de aprendizaje con refuerzo. La presente invención supera los problemas antes señalados a través de las características enumeradas en las reivindicaciones. In response to the above challenge, the invention proposes a virtualizable software service that can be dynamically instantiated at different points, end-to-end, of the multimedia content distribution, broadcast and / or transport network. The invention makes it possible to provide a quality measure, similar to what a user could provide, making use exclusively of information from the network and the content itself, without the need to rely on versions of that content prior to the processing stages that precede it. in the chain of distribution, diffusion and / or transport of the same. The quality measure offered by the system, provided as a numerical value within a limited scale, comes from an algorithm developed with the help of mathematical algorithms, signal processing and machine learning techniques, and has the ability to self-adapt through the information it is able to collect during its operation, relying on reinforcement learning techniques. The present invention overcomes the problems noted above through the features listed in the claims.
En particular, la invención permite, mediante un software compacto, instanciable como una máquina virtual o contenedor, la realización del cálculo de la predicción de calidad de un contenido audiovisual con distintas propiedades (resolución en el caso del vídeo, estándar de codificación tanto para audio como para vídeo, bits/muestra de audio, tasa binaria tanto de audio como de vídeo, etc.), ofrecido como un resultado numérico en una escala acotada fácilmente inteligible, a través de una interfaz visual, que puede igualmente ser almacenado y posteriormente recuperado de forma estructurada. In particular, the invention allows, by means of a compact software, instantiable as a virtual machine or container, the performance of the calculation of the quality prediction of an audiovisual content with different properties (resolution in the case of video, encoding standard for both audio as for video, bits / audio sample, bit rate for both audio and video, etc.), offered as a numerical result on an easily intelligible scale, through a visual interface, which can also be stored and later retrieved in a structured way.
Comparada con otras propuestas del estado de la técnica, esta propuesta tiene: 1) capacidad de funcionar de forma virtualizada, permitiéndose su instanciación de forma dinámica en distintos puntos y con distintas duraciones temporales a través de distintas soluciones técnicas de virtualización; 2) capacidad para trabajar sobre distintos contenidos en paralelo, o sobre el mismo contenido en distintos puntos de la red de distribución, difusión y/o transporte de contenidos audiovisuales; 3) capacidad de capturar y procesar los datos resultantes de su ejecución para la mejora de sus propios modelos de medida, aplicando técnicas de aprendizaje máquina por refuerzo; 4) capacidad de ofrecer los resultados en escalas ajustables e, incluso, definidas por el usuario, atendiendo a diferentes sensibilidades a capacidades de discriminación frente a posibles defectos presentes en el contenido; 5) capacidad de establecer distintas estructuras de almacenamiento para los datos, trabajar con ellos de forma individual y agregada, y recuperarlos según diferentes criterios fijados por el usuario. Compared with other state-of-the-art proposals, this proposal has: 1) the capacity to function in a virtualized way, allowing it to be dynamically instantiated at different points and with different time durations through different virtualization technical solutions; 2) ability to work on different content in parallel, or on the same content at different points in the distribution, broadcast and / or transport network of audiovisual content; 3) ability to capture and process the data resulting from its execution to improve its own measurement models, applying machine learning techniques by reinforcement; 4) ability to offer results in adjustable scales and even user-defined, taking into account different sensitivities to discrimination capabilities against possible defects present in the content; 5) ability to establish different storage structures for data, work with them individually and in aggregate, and retrieve them according to different criteria set by the user.
No es objeto de la presente invención ninguna reivindicación sobre el sistema de almacenamiento estructurado de la información, ni las tecnologías de presentación y visualización de datos ligadas a la interfaz gráfica. The object of the present invention is not any claim on the structured information storage system, nor the data presentation and visualization technologies linked to the graphical interface.
Aplicación industrial Industrial application
Las posibilidades de aplicación son múltiples, ya que la invención ofrece numerosas ventajas. El sistema permite la evaluación de calidad de contenidos audiovisuales ofreciendo un resultado susceptible de ser usado para: • Proporcionar un valor en una escala de calidad que permita la calificación y discriminación de servicios de contenidos audiovisuales independientemente de su naturaleza y plataforma de distribución, difusión o transporte. The application possibilities are manifold, since the invention offers numerous advantages. The system allows the evaluation of the quality of audiovisual content, offering a result that can be used for: • Provide a value on a quality scale that allows the qualification and discrimination of audiovisual content services regardless of their nature and distribution, broadcast or transport platform.
• Proporcionar sugerencias visuales y confiables para los proveedores de servicios audiovisuales en red que les permitan garantizar la calidad de la experiencia de los usuarios. • Provide visual and reliable suggestions for network audiovisual service providers that allow them to guarantee the quality of the user experience.
• Proporcionar alarmas e indicadores ante situaciones de calidad crítica en servicios de distribución, difusión y/o transporte de contenidos audiovisuales que permitan la toma de decisiones sobre la reconfiguración de los mismos. • Provide alarms and indicators in situations of critical quality in audiovisual content distribution, broadcast and / or transport services that allow decision-making on their reconfiguration.
Todo lo anterior abre la posibilidad de uso del sistema en: All of the above opens the possibility of using the system in:
• Verificación de calidad de señales audiovisuales de contribución y distribución.• Quality verification of audiovisual contribution and distribution signals.
• Análisis de rendimiento de modelos de codificación y multiplexación estadística, aplicada a múltiplex terrestres o segmentos espectrales satelitales. • Performance analysis of coding and statistical multiplexing models, applied to terrestrial multiplexes or satellite spectral segments.
• Análisis de calidad de servicios y programas de media en red y verificación de cumplimiento de las políticas de calidad de los proveedores de servicios audiovisuales. • Analysis of the quality of services and network media programs and verification of compliance with the quality policies of audiovisual service providers.
• Determinación del impacto de las configuraciones técnicas y de servicio sobre la calidad audiovisual percibida. • Determination of the impact of technical and service configurations on perceived audiovisual quality.
• Cuantificación del impacto sobre los contenidos audiovisuales de distintas plataformas de distribución, difusión y/o transporte para proveedores y agregadores de contenidos. • Quantification of the impact on the audiovisual content of different distribution, broadcast and / or transport platforms for providers and content aggregators.
De modo general, la invención permite la realización de cualquier aplicación de monitorización y análisis de calidad de contenidos audiovisuales y servicios de contenidos audiovisuales. In general, the invention allows the realization of any application for monitoring and analyzing the quality of audiovisual content and audiovisual content services.

Claims

REIVINDICACIONES
1. Método para la estimación de calidad percibida en señal audiovisual implementado por ordenador, donde la señal audiovisual se transmite por al menos una red de comunicaciones con contenidos audiovisuales seleccionada entre red de distribución, red de difusión y red de transporte; el método comprende los siguientes pasos: 1. Method for estimating perceived quality in audiovisual signal implemented by computer, where the audiovisual signal is transmitted by at least one communication network with audiovisual content selected from the distribution network, broadcast network and transport network; the method comprises the following steps:
- capturar unos parámetros característicos del servicio de transporte de datos con independencia de la red, tanto para la señal de vídeo como la señal de audio contenidas en la señal audiovisual (1); - capture some characteristic parameters of the data transport service regardless of the network, both for the video signal and the audio signal contained in the audiovisual signal (1);
- calcular un conjunto de métricas sin referencia, a partir de los parámetros característicos del servicio de transporte de datos, que proporcionan un conjunto ordenado de parámetros con los niveles de distorsión de calidad de los contenidos audiovisuales (2); y, - calculate a set of metrics without reference, from the characteristic parameters of the data transport service, which provide an ordered set of parameters with the quality distortion levels of the audiovisual content (2); Y,
- obtener un valor de calidad percibida aplicando algoritmia matemática, procesado de señal y técnicas de aprendizaje máquina sobre los parámetros característicos del servicio y el conjunto de métricas sin referencia (3). - Obtain a perceived quality value by applying mathematical algorithms, signal processing and machine learning techniques on the characteristic parameters of the service and the set of metrics without reference (3).
2. Método para la estimación de calidad percibida en señal audiovisual implementado por ordenador, según la reivindicación 1 , caracterizado porque los parámetros característicos del servicio (13) están seleccionados entre la resolución de la imagen, el número de canales y la frecuencia de muestreo del audio, las tasas binarias empleadas en la codificación de audio y de vídeo, y combinaciones de los anteriores. 2. Method for estimating perceived quality in audiovisual signal implemented by computer, according to claim 1, characterized in that the characteristic parameters of the service (13) are selected from the resolution of the image, the number of channels and the sampling frequency of the audio, the bit rates used in audio and video encoding, and combinations of the above.
3. Método para la estimación de calidad percibida en señal audiovisual implementado por ordenador, según la reivindicación 1, caracterizado porque calcular un conjunto de métricas sin referencia (21) adicionalmente comprende calcular una distribución de energía en función de la frecuencia y detectar unas estructuras tanto en la señal de audio como en la señal de video con capacidad de producir incomodidad perceptual procedentes de la codificación. 3. Method for estimating perceived quality in audiovisual signal implemented by computer, according to claim 1, characterized in that calculating a set of metrics without reference (21) additionally comprises calculating an energy distribution as a function of frequency and detecting structures both in the audio signal as in the video signal with the ability to produce perceptual discomfort from encoding.
4. Método para la estimación de calidad percibida en señal audiovisual implementado por ordenador, según la reivindicación 1, caracterizado porque la algoritmia matemática, el procesado de señal y las técnicas de aprendizaje máquina son regresores de Lasso y Ridge, regresores con máquinas de soporte “SVR”, Random Forest, y unos modelos de Deep Learning (3). 4. Method for estimating perceived quality in audiovisual signal implemented by computer, according to claim 1, characterized in that the mathematical algorithm, the signal processing and the machine learning techniques are Lasso and Ridge regressors, regressors with support machines " SVR ”, Random Forest, and some Deep Learning models (3).
5. Sistema para la estimación de calidad percibida en señal audiovisual, la cual se transmite por al menos una red de comunicaciones con contenidos audiovisuales seleccionada entre red de distribución, red de difusión y red de transporte; estando el sistema desplegado en una máquina virtual, dicho sistema está caracterizado porque comprende: 5. System for estimating perceived quality in audiovisual signal, which is transmitted by at least one communication network with audiovisual content selected from distribution network, broadcast network and transport network; the system being deployed in a virtual machine, said system is characterized in that it comprises:
• un módulo de captura configurado para capturar unos parámetros característicos del servicio de transporte de datos con independencia de la red, tanto para la señal de vídeo como la señal de audio contenidas en la señal audiovisual (1);• a capture module configured to capture some characteristic parameters of the data transport service regardless of the network, both for the video signal and the audio signal contained in the audiovisual signal (1);
• un módulo de cálculo de métricas de calidad sin referencia configurado para, a partir de los parámetros característicos del servicio de transporte, proporcionar un conjunto ordenado de parámetros con los niveles de distorsión de calidad de los contenidos audiovisuales (2); y, • a module for calculating quality metrics without reference configured to, based on the characteristic parameters of the transport service, provide an ordered set of parameters with the quality distortion levels of audiovisual content (2); Y,
• un módulo de predicción de calidad configurado para obtener un valor de calidad percibida aplicando algoritmia matemática, procesado de señal y técnicas de aprendizaje máquina sobre los parámetros característicos del servicio y el conjunto de métricas sin referencia (3). • a quality prediction module configured to obtain a perceived quality value by applying mathematical algorithms, signal processing and machine learning techniques on the characteristic parameters of the service and the set of metrics without reference (3).
6. Programa de ordenador que comprende instrucciones que, al ejecutar el programa en un ordenador, hacen que el ordenador lleve a cabo las etapas del método para la estimación de calidad percibida en señal audiovisual implementado por ordenador de cualquiera de las reivindicaciones 1 a 4. 6. Computer program comprising instructions that, when executing the program on a computer, cause the computer to carry out the steps of the method for estimating perceived quality in computer-implemented audiovisual signal of any of claims 1 to 4.
7. Medio legible por ordenador que contiene el programa de ordenador de la reivindicación 6 y que al ser leído y ejecutado por un ordenador pone en práctica el método reivindicado en cualquiera de las reivindicaciones 1 a 4. 7. Computer-readable medium that contains the computer program of claim 6 and that when read and executed by a computer implements the method claimed in any of claims 1 to 4.
PCT/ES2020/070809 2020-02-07 2020-12-21 Method and system for estimating perceived quality in an audiovisual signal WO2021156525A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
ES202030100A ES2767475A1 (en) 2020-02-07 2020-02-07 METHOD AND SYSTEM FOR THE ESTIMATION OF PERCEIVED QUALITY IN AUDIOVISUAL SIGNAL (Machine-translation by Google Translate, not legally binding)
ESP202030100 2020-02-07

Publications (1)

Publication Number Publication Date
WO2021156525A1 true WO2021156525A1 (en) 2021-08-12

Family

ID=71080015

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/ES2020/070809 WO2021156525A1 (en) 2020-02-07 2020-12-21 Method and system for estimating perceived quality in an audiovisual signal

Country Status (2)

Country Link
ES (1) ES2767475A1 (en)
WO (1) WO2021156525A1 (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
ES2968713A1 (en) * 2023-12-04 2024-05-13 Univ Madrid Politecnica SYSTEM AND METHOD FOR ESTIMATING PERCEIVED QUALITY AND ITS IMPACT ON THE CONSUMER OF MULTIMEDIA AUDIOVISUAL CONTENT (Machine-translation by Google Translate, not legally binding)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2808829A1 (en) * 2013-05-28 2014-12-03 Vrije Universiteit Brussel Method and apparatus for predicting the perceptual quality of a signal
CN108900862A (en) 2018-02-14 2018-11-27 南京邮电大学 Network video stream QoE-QoS parameter mapping method based on statistical analysis
US20190124375A1 (en) 2016-06-29 2019-04-25 Telefonaktiebolaget Lm Ericsson (Publ) Quality Estimation Of Adaptive Multimedia Streaming

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2014209493A1 (en) * 2013-06-28 2014-12-31 Wi-Lan Labs, Inc. State management for video streaming quality of experience degradation control and recovery using a video quality metric
KR102137730B1 (en) * 2014-04-17 2020-07-24 삼성전자주식회사 QoE Provisioning Method And QoE Provisioning Apparatus for Mobile Video Application
JP7072754B2 (en) * 2016-03-06 2022-05-23 エスシムウェーブ インク． Methods and systems for automatic quality measurement of streaming video

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2808829A1 (en) * 2013-05-28 2014-12-03 Vrije Universiteit Brussel Method and apparatus for predicting the perceptual quality of a signal
US20190124375A1 (en) 2016-06-29 2019-04-25 Telefonaktiebolaget Lm Ericsson (Publ) Quality Estimation Of Adaptive Multimedia Streaming
CN108900862A (en) 2018-02-14 2018-11-27 南京邮电大学 Network video stream QoE-QoS parameter mapping method based on statistical analysis

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
BRANDAO ET AL: "No-reference image quality assessment based on DCT domain statistics", SIGNAL PROCESSING, ELSEVIER SCIENCE PUBLISHERS B.V. AMSTERDAM, NL, vol. 88, no. 4, 8 January 2008 (2008-01-08), pages 822 - 833, XP022412980, ISSN: 0165-1684, DOI: 10.1016/J.SIGPRO.2007.09.017 *
MUHAMMAD SHAHID ET AL: "A reduced complexity no-reference artificial neural network based video quality predictor", IMAGE AND SIGNAL PROCESSING (CISP), 2011 4TH INTERNATIONAL CONGRESS ON, IEEE, 15 October 2011 (2011-10-15), pages 517 - 521, XP032070622, ISBN: 978-1-4244-9304-3, DOI: 10.1109/CISP.2011.6099931 *
XIUHUA JIANG ET AL: "No-Reference Perceptual Video Quality Measurement for High Definition Videos Based on an Artificial Neural Network", COMPUTER AND ELECTRICAL ENGINEERING, 2008. ICCEE 2008. INTERNATIONAL CONFERENCE ON, IEEE, PISCATAWAY, NJ, USA, 20 December 2008 (2008-12-20), pages 424 - 427, XP031403397, ISBN: 978-0-7695-3504-3 *
ZHENGYOU WANG ET AL: "No-reference image quality assessment for compressed images based on DCT coefficient distribution and PSNR estimation", SIGNAL&INFORMATION PROCESSING ASSOCIATION ANNUAL SUMMIT AND CONFERENCE (APSIPA ASC), 2012 ASIA-PACIFIC, IEEE, 3 December 2012 (2012-12-03), pages 1 - 4, XP032309962, ISBN: 978-1-4673-4863-8 *

Also Published As

Publication number Publication date
ES2767475A1 (en) 2020-06-17

Similar Documents

Publication Publication Date Title
Nightingale et al. 5G-QoE: QoE modelling for ultra-HD video streaming in 5G networks
ES2653462T3 (en) Concept for determining the quality of a multimedia data stream with a variable bit rate
KR100922898B1 (en) QoE guaranteed realtime IP-media video quality measurement apparatus and measurement method thereof
US8027264B2 (en) Method and apparatus for measuring subjective assessment of digital video impairment
Yang et al. Bitstream-based quality assessment for networked video: a review
Leszczuk et al. Recent developments in visual quality monitoring by key performance indicators
US7020093B2 (en) Delivery of streaming media
Yang et al. Content-adaptive packet-layer model for quality assessment of networked video services
Paudyal et al. A study on the effects of quality of service parameters on perceived video quality
US9077972B2 (en) Method and apparatus for assessing the quality of a video signal during encoding or compressing of the video signal
US20220214545A1 (en) Method and system for filtering a panoramic video signal
WO2021156525A1 (en) Method and system for estimating perceived quality in an audiovisual signal
Suárez et al. Assessing the qoe in video services over lossy networks
Barkowsky et al. Hybrid video quality prediction: reviewing video quality measurement for widening application scope
Wang et al. No-reference hybrid video quality assessment based on partial least squares regression
Anegekuh et al. Encoding and video content based HEVC video quality prediction
Pullano et al. PSNR evaluation and alignment recovery for mobile satellite video broadcasting
Kim et al. Subjective and objective quality assessment of videos in error-prone network environments
Reiter et al. Comparing apples and oranges: assessment of the relative video quality in the presence of different types of distortions
Babić et al. Real-time audio and video artifacts detection tool
WO2018064934A1 (en) Method and device for evaluating video quality
Mu et al. Discrete quality assessment in IPTV content distribution networks
Pal et al. Model for mobile online video viewed on samsung galaxy note 5
Martínez-Rach et al. On the performance of video quality assessment metrics under different compression and packet loss scenarios
Llorente et al. Efficient Computational Cost Saving in Video Processing for QoE Estimation

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20845162

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20845162

Country of ref document: EP

Kind code of ref document: A1