Search

A face presentation attack detection (face-PAD) system is in charge of determining whether a face corresponds to a presentation attack or not. The vast majority of proposed solutions consider a static scenario, where models are trained and evaluated in datasets where all types of attacks and conditions are known beforehand. However, in a real-world scenario, the situation is very different. There, for instance, the types of attacks change over time, with new impersonation situations appearing for which little training data is available. In this paper we propose to tackle these problems presenting for the first time a con-tinuallearning framework for PAD. We introduce a continual meta-learning PAD solution that can be trained on new attack scenarios, following the continual few-shot learning paradigm, where the model uses only a small number of training samples. We also provide a thorough experimental evaluation using the GRAD-GPAD benchmark. Our results confirm the benefits of applying a continual meta-learning model to the real-world PAD scenario. Interestingly, the accuracy of our solution, which is continuously trained, where data from new attacks arrive sequentially, is capable of recovering the accuracy achieved by a traditional solution that has all the data from all possible attacks from the beginning. In addition, our experiments show that when these traditional PAD solutions are trained on new attacks, using a standard fine-tuning process, they suffer from catastrophic forgetting while our model does not.

When assessing product life, the survival analysis is generally conducted in a time or usage domain. In certain instances, it is beneficial to investigate the impact of joint variables on product reliability, and in this case, the joint distribution of usage and time are considered. While it is simple to analyze unconditional probability density functions independently, the problem of statistical independence with random variables of usage and time arises. Usage is not independent of time, since usage contains information about time and thus, the joint probability density function then cannot be the product of both marginal probability density functions.

Self-consumption is a growing public demand in an energy environment with growing electricity costs and decreasing photovoltaic installation costs. Shared self-consumption is an imperative aspect for bringing self-consumption into Multi-Family Residential Buildings (MRB), where most families live. Nevertheless, current legislation in most countries does not consider shared self-consumption or does not exploit its full potential; such is the case of Spain or Portugal. This paper will present a novel optimization problem for studying the economics of a shared self-consumption installation in a MRB (composed of five family demands, a PV installation and a battery) with the aim of reducing the total bill of the MRB during an entire year. The impact on energy communities of two different types of energy policies is analysed: the remuneration scheme for the surplus energy (net metering, net billing, and exclusive self-consumption policies) and the regulation for shared self-generated energy (demand-dependent, proportional output and no sharing). It is found that the regulation for the sharing energy can be more important that the remuneration scheme, which has been the traditional target of the self-consumption policy.

The increase in popularity of video streaming, the improvement of bandwidth corresponding to 5G networks and the transmission of higher amounts of data derived from advanced video formats such as Ultra High Definition (UHD) with 4K and 8K resolutions make the user demand a high perceptual quality of the contents consumed. For that reason, it is necessary to create advanced models for predicting video quality based on the video features and the encoding settings in environments where the reference is not present. The use of machine learning (ML) techniques for data analysis based on patterns extracted on the features of audio-visual content, improves the generation of prediction models for accurately predicting quality. This paper presents a novel model for assessing video quality based on the analysis of encoding video settings of the transmitted contents and the image intrinsic characteristics for objectively estimating the Mean Opinion Score (MOS) in correlation with the subjective results. The use of data mining techniques for combining a collection of parameters associated to the video transmitted improves the performance of traditional quality evaluation, as demonstrated with the database analyzed for this purpose.

Recommendations such as P.910 suggests parameters TI (temporal information) and SI (spatial information) for characterizing video sequences for quality assessment. In this paper, we suggest two additional parameter based on disparity called SPI (spatial parallax information) and TPI (temporal parallax information) to characterize 3DTV video sequences for this purpose.

Disparity may cause visual discomfort. Pairs of video sequences with different levels of parallax, both negative and positive, were presented together to the observers. The observers evaluated the cases in which visual discomfort occurred after visualizing the transition on each pair.

Assessing video quality is a complex task. While most pixel-based metrics do not present enough correlation between objective and subjective results, algorithms need to correspond to human perception when analyzing quality in a video sequence. For analyzing the perceived quality derived from concrete video artifacts in determined region of interest we present a novel methodology for generating test sequences which allow the analysis of impact of each individual distortion. Through results obtained after subjective assessment it is possible to create psychovisual models based on weighting pixels belonging to different regions of interest distributed by color, position, motion or content. Interesting results are obtained in subjective assessment which demonstrates the necessity of new metrics adapted to human visual system.

We have developed a Real Time Kinematic (RTK) navigation system that uses pseudolite signals in addition to the GNSS satellite signals. This RTK+ system can be used under shadowed conditions where satellite signals alone do not provide satisfactory availability or performance. Such conditions exist for example under a large harbor crane, where continuous positioning of straddle carriers handling the containers would be very useful for logistics. Single-frequency test results of RTK+ show decimeter-level accuracy and rapid initialization of carrier ambiguities when the receiver moves. The results are considerably better than could be obtained with single-frequency RTK using satellite signals only. European Conference of Postal and Telecommunications Administrations (CEPT) is preparing a regulatory framework for the use of outdoor pseudolites. It is expected that it will open the way for commercial deployment of RTK+.

This paper presents a solution for inserting graphics in a 3D HDTV environment in real-time. It focuses the effort on optimizing the process of insertion in order to override depth cue conflicts between occlusion and stereoscopical disparity caused by the insertion, thus, providing a method to control the 3D QoE after the insertion. The authors propose an algorithm based on disparity estimation modified to meet the needs of this particular task. Based on Census transform disparity estimation, the introduction of reduced disparity maps and mode filters offers the possibility to increase the disparity range and achieve subjective visual improvements while keeping the amount of physical resources needed. Visual discomfort avoidance versus objective stereoscopical analysis performance is a key point of the study, because it enables the possibility of overriding heavy algorithms keeping the quality of experience at a high level. It is intended to add On Screen Displays, device or channel logos, advertising or even subtitles to a 3D HDTV workflow on a general hardware device, meaning that it can even fit on private electronic devices, not only on professional equipment.

This paper contains selected test results of an indoor positioning prototype system making use of GPS/Galileo signals on L1/E1 and L5/E5a in addition to WiFi data, ZigBee data and a MEMS-based dead reckoning algorithm called pedestrian navigation system (PNS). The system uses a partially coherent ultra-tightly coupled integration scheme. The GNSS signal processing (acquisition and tracking) supports coherent integration times up to several seconds, which reduces the squaring loss and mitigates multipath by the high Doppler selectivity of a few hertz. The system works in post processing and is intended to be an indoor analysis system allowing to get a better understanding of the indoor positioning problem and to act as a reference system. GNSS signal acquisition is demonstrated at ∼ 3 dBHz and tracking well below 0 dBHz.

The goal of this paper is to present the architecture and first performance results of the DINGPOS platform. The DINGPOS platform (Demonstrator for INdoor GNSS POSitioning) is a project funded by the ESA that covers the design, development and integration of an experimental indoor positioning system based on the fusion of three different technologies: High Sensitivity GNSS (GPS and the future Galileo), MEMS-based Pedestrian Navigation System and WIFI. This paper introduces the different architectural trade-offs of the final platform and presents the first results of its performance and capabilities with real data.

Beta-barrel membrane proteins (MP) are found in Gram-negative bacteria, mitochondria and chloroplasts. They play important roles in metabolism of bacteria, where they are involved in transport of solutes in and out of the cell. Beta-barrel proteins may also act as proteases, lipases and may be important for cell-cell adhesion. Currently, there are about 30 non-redundant solved structures of beta-barrels. Although the number of b-barrel folds is fairly small, it is possible to expand the amount of available structural information by homology modeling using existing structures as templates. The scope of structure prediction may be widened by finding remote homologues of the existing structures. To improve the sensitivity of the database searches and the quality of sequence alignments, we first study evolutionary history of transmembrane segments of 7 beta-barrel membrane proteins by estimating substitution rates with a Bayesian Monte Carlo approach. Next, we calculate amino acid substitution matrices, beta-barrel Transmembrane scoring Matrices (bbTM), specifically tuned for TM regions, which can be used to detect remote homologues. We then test bbTM matrices by comparing their performance with membrane-protein derived scoring matrices PHAT and SLIM. Our results demonstrate that bbTM matrices have higher selectivity towards transmembrane beta-barrel proteins and may be used with higher confidence in database searches for remote homologues of this class of proteins.

Media environment is nowadays evolving dynamically integrating a more active participation of users. In this context, Media campus consists of the first Spanish experience dealing with content delivery from high definition adapted terminals to mobile handheld devices, with both professional and user generated audiovisual contents. Media includes advanced and accessible interactivity. Media campus uses DVB-T/H hierarchy transmission along with AVC/H.264 coding to provide television services within Ciudad Universitaria area in Madrid. The delivery platform is composed by a Web portal that allows user generated media monitoring and exchanging, and a digital television head-end. User contributions are processed and adapted, and then the multimedia content of high interest is broadcasted to TV and mobile devices following DVB standards. From the experience new models and algorithms to control several aspects of the service will be test; automatic quality of service adjustment, consumption metrics and personalization of contents.

Although in the past decades bandwidth capacities have increased in existing networks, video compression techniques are still essential for managing high bit rates required by video services, especially those related with the trends on high definition video services. Using compression implies in most cases lossy processes that, in addition to reduce average video quality, create undesirable errors called artifacts. The treatment and removal of these artifacts from video sequences is a primary goal in nowadays coding standards. One of the most common and perceptually annoying artifacts is called tilling. Unwillingly produced by many digital coding systems, such as H.264, its particular treatment is a new investigation line that is described deeply in this paper.

Proceeding of: 2016 European Conference on Networks and Communications (EuCNC) 27-30 junio 2016 Athens, Greece A transport network for 5G envisions integrating the fronthaul and backhaul segments (namely 5G-Crosshaul) into a single transport network. This requires a fully integrated and unified management of fronthaul and backhaul resources in a sharable, scalable and flexible way. The integrated and unified management of the transport network resources follows the Software Defined Networking (SDN) principles of centralization, decoupling between control and data plane and application ecosystem. Innovation to enable context-aware resource management and to provide system-wide optimization of QoS, as well as energy related objectives, lies on defining key applications as logical decision entities to program the underlying network and packet forwarding behaviour. This paper presents a set of 5G-Crosshaul SDN applications with the scope of (i) managing the 5G-Crosshaul resources, including not only network but also computing and storage resources, and (ii) offering efficient media distribution and traffic offloading solutions. In order to develop these applications in a real system, we identify the interactions between the applications and the control plane. Based on these interactions, we propose a two-layer application plane and specify the requirements of the applications on the Northbound Interface (NBI) of the control plane. This work has been supported by the H2020 project “5G-Crosshaul: The 5G Integrated fronthaul/backhaul” (671598).