Partner Description
|
VISNET demonstrates the formation of a critical mass by putting together 15 leading organisations across Europe in the field of networked AV systems and home platforms. The consortium consists of 9 universities, 3 research institutes and 3 industrial Telecom companies, 2 of which are major network operators in their respective states. All members are known for their proven track record in audiovisual information and networking technologies and enjoy a national and international reputation in the field. VISNET integrates a number of researchers who have made significant contributions to the advance of this field through standardisation activities, books and paper publications, patents and spin-off SMEs, as well as many other academic and industrial achievements. The table below shows the participants acronyms as well as their respective numbers used in this proposal. PartnersVISNET has 15 partners in 9 countries and integrates over 100 researchers
University of Surrey, UKMultimedia Communications Research Group, Centre for Communication Systems Research (CCSR)Visnet will be coordinated by CCSR, a part of an academic school at the University of Surrey which is one of the two only universities in the UK which have achieved a 5*A ranking (highest research ranking given to an academic university in the UK) in the last RAE (Research Assessment Exercise, 2001) in the Electronic Engineering discipline. CCSR is a large academic research centre with over 100 personnel divided into 3 main research groups, one of which is the multimedia communication research group. The Multimedia group is largely involved in research activities covering speech, image and video compression algorithms, AV transmissions over networks, media adaptation for AV transmissions over heterogeneous networks, QoS control and security. The group also contributes extensively to the joint programme of activities of Visnet and makes use of its new high-end lab facilities (I-lab) that consists of a multi-disciplinary mixed reality media lab environment. The I-lab facilities along with the campus-wide wireless network testbed (WNT) are used in several activities related to the creation of 3D AV material as well as the performance evaluation of the proposed coding and transcoding algorithms under real network conditions. The group has also experience in the set-up of spin-out SME companies. This experience will be vital for supporting the dissemination activities and durability plans beyond the budgeted period of the Commission. Fraunhofer Institute for Telecommunications, Heinrich-Hertz Institute, GermanyImage Processing DepartmentThe Fraunhofer Institute for Telecommunications, Heinrich-Hertz-Institute (HHI) is a German research institute in the field of communications, located in Berlin. Our activities are financed by both institutional funding from the Federal Government and the State of Berlin, and external funding from industry and research organisations. About 280 people work on solutions to problems related image processing, multimedia, human factors, photonic networks, and mobile broadband communication. The focus of the Image Processing Department with more than 50 employees is on digital signal, image, and video processing. In more than 20 research projects, new algorithms and applications are developed, ranging from very low bit-rate video coding for mobile and Internet applications, immersive video conferencing, video surveillance, face tracking and mimic analysis, and 3D-TV. In the past, the HHI participated in several EUREKA projects (EU95, ADDT, VADIS, MEDEA), RACE-projects (FLASH, TRANSIT, DISTIMA, HAMLET), ACTS projects (CINENET, EMPHASIS, PANORAMA, MOMUSYS, CustomTV), IST-projects (DRIVE, SAMBITS), and the ESPRIT project NEMESIS. Currently, the HHI is member of the IST projects ATTEST, MASCOT and VIRTUE. Nationally, the HHI was leading the joint BMBF research projects HDTV-T, MINT, and INVINET and is currently involved in BBI, KOKON, ITI, and VisionIC. The HHI is a competence center for video coding and processing which is reflected in leading activities and contributions to standardisation efforts of the ITU-T and MPEG. Substantial know-how also exists in the fields of 3D signal processing, image analysis and synthesis, computer vision, and porting of signal processing algorithms to hardware architectures. Especially the work on facial expression analysis and animation, 3D object geometry estimation, image-based rendering, and coding and streaming of image sequences can beneficially be integrated into the VISNET activities. Ecole Polytechnique Fédérale de Lausanne, SwitzerlandThe TV group of Signal Processing Institute of EPFL will be involved in this network. The group is active in research and teaching in the field of visual information processing and coding, in many areas directly relevant to the objectives of the proposed network. TV group currently has 8 researchers, mostly pursuing Ph.D. theses, 4 post-doctoral researcher, and many undergraduate students. Undergraduate students are EPFL students or full-time visiting students from other universities, each staying for a period of around 6 months and involved in one of the research projects of the group. The research topics span through three highly interconnected disciplines of imaging, namely, compression, processing/interpretation, and security. Compression includes still image, video and 3D image coding. Processing/interpretation includes image and video analysis such as change detection, feature point extraction and tracking, object segmentation and tracking, compressed domain processing, and multi-modal man-machine interface for interactive communication. Security activity includes development of algorithms for copyright protection, conditional access and authentication of images, video as well as 3D models. The research activities are carried out in the framework of various projects coming from different sources. These sources include fundamental research grants from the Swiss authorities (FNRS- Swiss National Foundation for Scientific Research), Academia-industry cooperation program (CTI), European Projects (IST), and direct industrial contracts. In the last years, the team has been involved in more than a dozen of such frameworks. Current research projects include, one FNRS project: interactive video systems; five European projects: Interface on man-machine interaction through audio-visual analysis, Art.Live on new media creation and experience, Certimark on evaluation of watermarking techniques and new algorithms; Perseo on universal multimedia access, and 2Kan on JPEG 2000 advanced networking. In addition to the above, the team is very active in standardization activities MPEG-4, MPEG-7 and MPEG-21, as well as JPEG 2000, where it contributes both at technical and leadership levels. Instituto Superior Técnico, PortugalInstituto Superior Técnico (IST) is part of Technical University of Lisbon. The IST Image Group research interests cover the areas of visual information analysis, modeling, processing, manipulation, representation, coding and description. IST has developed, in the last years, intense research and development activities in image processing and coding, from still pictures to mobile videotelephony, digital TV, and advanced interactive systems. These activities, often developed in collaboration with international partners (universities, industries, telecommunication operators and research institutes), allowed the preparation of many M.Sc. and Ph.D. thesis, leading to a significant experience in multimedia processing and representation technologies. IST has been active in international research and standardisation groups, such as COST 211, ISO/JPEG, and mainly ISO/MPEG. IST was partner of RACE MAVT (which developed the first hardware demonstrator of a low bitrate mobile audiovisual terminal), ACTS MOMUSYS (which developed an MPEG-4 hardware demonstrator), and ACTS DICEMAN (which provided important inputs to MPEG-7 standardisation). IST was involved in launching the MPEG-4 and MPEG-7 standardisation efforts. IST participation in MPEG-4, focused in requirements and profiles specification, shape coding, error resilience, video rate control, 3D facial animation. IST was involved in several proposals that become the basis of MPEG-7 image and video description schemes. IST is also involved in the development of the MPEG-21 standard, especially in terms of usage environment description and scalable video coding. IST experts have chaired many MPEG Ad Hoc Groups and the Requirements Group. The interest of IST in the VISNET project follows its interest on networked AV technologies. More precisely, IST wil l bring to VISNET technical knowledge and experience to provide adequate solutions for: Video analysis, notably segmentation and evaluation of segmentation algorithms and results, video description for efficient retrieval, browsing, filtering and summarization, advanced video coding, notably targeting improved efficiency, scalability and error resilience, specially in view of improved mobile and Internet video communications. Due to its experience in standardisation bodies, IST is also determined to play a central role in the connection between MPEG and VISNET. IST has the necessary equipment and resources to fulfil with success the tasks it will be in charge in the context of the VISNET NoE. Aristotle University of Thessaloniki, GreeceArtificial Intelligence & Information Analysis Laboratory, Computer Vision and Image Processing Group, Department of InformaticsThe computer vision and digital image processing group at theDepartment of Informatics, Aristotle University of Thessaloniki is an active research group working for more than fifteen years in multimedia and digital video/image/signal processing and analysis as well as several related areas. Its topics of interest and current research span the areas of digital image/video processing and analysis, computer vision, watermarking and IPR protection, biometrics, 3D image processing, virtual reality, speech processing, statistical language engineering, multimedia indexing and retrieval, fusion from different modalities, etc. The group has strong cooperation with other Universities, research centres and industrial partners in the areas of its interests through projects mainly funded by the European community. It has participated in a number of projects dealing with topics adjacent to those that will be addressed within the framework of the VISNET proposal: WORLDSTUDIO (3D graphics/VR, IST, EU), HYPERGEO (Mobile tourist geographical information systems, IST, EU), CAROUSSO (3D audio/video environments, IST, EU), MOUMIR (multimedia retrieval, RTN, IST, EU), MUHCI (multimodal human-computer interaction, RTN, EU), PENED99 (3D medical imaging, GSRT, Greece), CERTIMARK (benchmarking of watermark methods, IST, EU), INSPECT (Chaotic signal processing for watermarking, LTR, EU), OKAPI (Access control for multimedia systems, ACTS, EU), M2VTS (Multimodal person verification, ACTS, EU), ACCOPI (Access control and copyright protection RACE, EU), DISTIMA (Stereoscopic HDTV, RACE, EU), NOBLESSE (Nonlinear image analysis, LTR, EU), NAT (Nonlinear and adaptive techniques in digital image processing, analysis and computer vision, ESPRIT, EU), MAVI (Model based analysis of video information, HCM, EU), Image sequence processing and analysis (GSRT, Greece), Parallel color image processing and multi-channel signal processing (GSRT, Greece). As a result of this research, more than 420 publications in international scientific journals, and conference proceedings, 16 book chapters and 7 books have been published. Within VISNET, AUTH will deal with topics that include shot cut detection, face detection and tracking, speech analysis and watermarking. Through its involvement in the Joint Program of Activities, AUTH will benefit from the exchange of research results, joint research and the sharing of tools, databases, common platforms etc that will be created within the network AUTH is workpackage leader in WP4.2.2 "Semantic Video Segmentation and Tracking Universitat Politecnica de Catalunya, SpainAudio Visual Technologies GroupThe members of the Group of Audio Visual Technologies of the Universitat Politécnica de Catalunya, UPC, have been working for more than 15 years in audio and video coding, image and video analysis and mobile communications. Members of the group have developed in particular a number of low-level tools for image segmentation, object tracking and motion estimation. More recently, the group has focussed on video indexing, face detection and recognition and metadata based-coding. This research activity has been the basis for collaborations within international projects and for contributions to MPEG-4. Members of the group have been responsible of the organization of the International Conference on Image Processing, ICIP, which was held in Barcelona in September 2003. With regards to European projects, a member of the group has been the prime contractor of the RACE II MORPHECO project (development of region-based codec based on mathematical morphology). Moreover, there has been a strong contribution to IST project VIDAS (video codec assisted by audio information), IST project HYPERMEDIA (broadcast service of MPEG-2 sequences supported by MPEG), and IST project MASCOT "Metadata for Advanced Scalable Video Coding Tools". The experience and technical inputs provided in these projects, along with many related publications constitute a good background for our work in VISNET regarding new paradigms in video coding using Wyner-Ziv approaches, the use of metadata information to improve coding, and the fusion of multimodal information for face detection and recognition. Joint Research Centre, ItalyThe Joint Research Centre (JRC) is the European Commission's own research centre. IPSC, the Institute for the Protection and Security of the Citizen, is one of the institutes of the JRC, is located at Ispra, Italy and has a permanent staff of about 350 people. The SIR sector (Surveillance and Information Retrieval) within the NPNS Unit (Non-Proliferation and Nuclear Safeguards) is involved in research in the VISNET related areas: 3-D reconstruction of small objects and environments "as-built" and Internet-based control, integration and visualisation of remote sensory data. Research is done within the framework of the Nuclear Safeguards programme and is applied to the Remote Verification of Fissile Material Storage Areas. The sector has very good links with many European Universities, and hosted many researchers and students doing final year projects, Master and PhD research work. The JRC-Ispra provides an intellectually stimulating international and multi-disciplinary environment, with a highly qualified staff of about 1800 people from more than 30 nationalities. NPNS-SIR's laboratories have the equipment and software tools to support research projects including the VISNET network. The main contribution to VISNET is the research related to the creation of 3D photo-realistic, dimensionally accurate, reconstructed models from the reality "as-built". These models range from very small objects (centimetre scale) to buildings (tens of meters scale) and large scenes (kilometre scale). Such models can be used as the spatial framework to embed all other types of dynamic and static visual and audio data. Some examples of specialist equipment are:
NPNS-SIR has been (is) involved in the following European projects and research networks: RESOLV (DGXIII-ACTS) project aiming at the creation of photo-realistic 3-D reconstructed models of the reality "as-built" [JRC's role was to acquire, and develop all the 3-D-reconstruction algorithms, as well as to investigate new methods and applications for 3-D reconstruction]; INFOBOY (DGXIII-ESPRIT) 3-D technologies for Internet tourist advertisement of farmhouse holidays in Bavaria; ABI (DG-XIII ESPRIT) Long Term Research aiming at developing a brain actuated interface (e.g., mouse); SMART2 and CAMERA (DG RESEARCH TMR) research networks, dealing with robotics and 3D technologies respectively; TEST (DGXIII-ESPRIT) aiming at Tele-presence for Entertainment, Surveillance and Training using Virtual Reality and real-time Internet standards; NESTOR (DG ENTR) aiming at developing an autonomous mobile courier for transporting small goods inside hospitals and health centres. Thales Research and Technology, UKNetwork Systems GroupThales Research & Technology (UK) [TRT (UK)] is the UK arm of the Thales corporate research centre and provides close support to the various Thales operating companies. Thales designs, develops and implements advanced electronic solutions, which meet the communications, information and security of people in countries throughout the world. Its activities cover three major markets: Information Technology and Services, Aerospace and Defence. Thales has global revenues of €8.6b and employs 65,000 people, 12,000 of whom are employed in the UK. TRT (UK) employs approximately 120 staff of which some 80% are science, maths or engineering graduates. The activities within TRT(UK) are subdivided into 3 main areas; Navigation, Sensors & Signal Processing and Network Systems. The Network Systems Group specialises in technologies associated with modern communication networks, security and vision processing. The group has a network test bed facility, which can be used to generate the various network performance scenarios, and security architectures that will be needed for VISNET. Members of the Network Systems Group are also active participants in the UK’s Mobile Virtual Centre of Excellence (VCE), which sponsors university research into mobile radio systems, networks & security beyond 3G. Apart from internally funded programmes, members of the Networks Systems Group have, within the last 12 months, been working on the following domestic and European projects relevant to the activities embraced by VISNET proposal:
TRT(UK) will bring to VISNET expertise to the Joint Programme of Activity in the following areas: Networking, specifically for QoS in IP networks, security, specifically in security architectures and PKI, Image Processing, specifically in automating camera calibration systems and 3D-secene understanding. In addition to providing support for Thales business objectives, Thales Research & Technology (UK) offers a wide range of consultancy and development services to European Government Agencies and to industry throughout the world, examples include EC, NATO, ESA, INMARSAT, DERA, DTI, Aerospatiale, Alcatel, GMV and others. Since 1983, innovative products have resulted in Thales companies being awarded nine Queen’s Awards for Technological Achievement. For many of these products Thales Research & Technology (UK) has provided the key technology that has led to market success. Technical University Berlin, GermanyCommunications DepartmentTUB is one of the largest technical universities in Germany, home to excellent researchers in every area of science and technology. Two outstanding researchers head the communications department (Fachgebiet Nachrichtenuebertragung), Professors Peter Noll and Thomas Sikora. Whilst Prof. Noll is world renowned for his work in coding, compression and representation of audio signals and data, Prof. Sikora is equally famous for his contributions to video data compression and representation. Both have been heavily involved in numerous standardisation activities within e.g. ITU-T and ISO/IEC MPEG. Moreover, Prof. Sikora has been involved in several EU-FP5 projects. Currently, the communications department research covers many different topics connected to audiovisual communications. Segmentation, classification and interpretation of AV sensor data for use in immersive communications, security applications and others are carried out. Furthermore, automatic annotation and metadata generation schemes for AV data are being developed in combination with efficient search and data-mining schemes. Moreover, AV data compression and robust representation techniques are being investigated under numerous aspects such as lossless compression, transmission over hostile channels, joint source-channel coding. 18 researchers and PhD students work on these topics and are supported by 4 secretarial and technical staff. Vodafone Ltd, UKVodafone is one of the world's largest mobile telecommunications network companies. We have interests in mobile networks throughout the world and we serve over 112.5 million proportionate customers. The primary role of Vodafone Group Research and Development is seen as giving Vodafone the ability to deliver on it’s business strategy, which is heavily dependant on not only having an in-depth understanding of what is possible today but on having a clear vision of the future and what can effect its outcome. Current areas of research within R&D which are focussed on multimedia applications are: Evaluation of compression techniques (DCT, VQ, wavelet) and assessment of future video encoding technologies (e.g., H.26L), which will reduce bitrate requirements thereby improving efficiency and increasing capacity. Investigation of synthetic photo-realistic imaging technology, which can be used for realistic avatar-type applications, or replacing simple ‘head-and-shoulders’ video sequences using half the bandwidth. Demonstration of 2D and 3D vector graphics as a means of generating graphics on the device rather than downloading pre-rendered material encoded in one of the bitmap graphic formats and in the case of 3D allowing applications to become more immersive. Additionally, audio compression technologies and audio synthesis techniques in support of a range of applications, i.e. speech, background music and music delivery. Technologies that enable audio spatialisation (i.e., the generation of phantom audio sources), and the use of these techniques to enhance mono-aural signals, increase the immersiveness of games and assist speaker identification in video conferencing applications. Furthermore, implementation configurations of streaming protocols (collaboration with infrastructure vendors and streaming solution vendors; co-ordination within Standards (3GPP) to highlight deficiencies in the current PSS specifications and identify means of improving quality). Work also includes monitoring and participation (where appropriate) in the various Standards bodies and Industry fora involved in the definition, testing or licensing of media technologies, such as 3GPP, IETF, W3C, ITU-T, ISO MPEG, MMA, IMTC and JCP. Other activities include: Consideration of the Digital Rights Management requirements and implementation issues for the delivery of multimedia content, discussion with terminal and streaming solution providers of the options for testing and ensuring interoperability between PSS-compliant components from different suppliers (in conjunction with the Global Streaming Media Project) and tracking the status of MPEG-4 visual, AAC and H.26L licensing development. France Telecom R&D, FranceEurope's leading telecommunications R&D center, France Télécom R&D contributes to creating value for the Group. Its transversal, decentralized structure facilitates coordination between France Télécom's market strategy and the products developed. France Télécom's R&D activities underlie the Group development strategy based on stimulating growth in usage, Internet, mobiles, and international operations. By developing the services of tomorrow, France Télécom R&D contributes to creating value for the Group. Its missions include anticipating technological breakthroughs and radically new patterns of use. France Télécom R&D innovates to offer our customers the best in telecommunications today and in the future. Areas of research include human interactions (man-machine interfaces), mobility, network architecture, fixed/mobile/Internet convergence, and very high throughput transmission and access networks. And for the past fifty-five years, France Télécom R&D has been a major player in this world of telecommunications and information technologies. In its laboratories, some 3000 engineers, scientists, and technicians are working to develop the innovations which will soon be used by France Télécom's customers. Yet, before these products are actually used, how do they really work? What scientific and technical principles govern telephone service, Internet, access, transmission, higher bit rates, network infrastructure etc. Staff are located at 8 different sites in France, and a laboratory which was set up recently in California and Japan. France Télécom R&D is very actively running internal experiments in areas related to the scope of VISNET such as automated 3D modelling tools (laser or video based), using neural networks for still or moving 3D objects. FTRD has also vast expertise and contributions to new leading standards such as XML, X3D, MPEG4, and also with associated partners (Leading edge companies, and "Cultural Heritage" partners such as: Ministere de la Culture Paris, and Museums). Experiments have been done on outstanding objects and monuments such as: Mont Saint Michel, Arc de Triomphe, Avenue des Champs Elysées, Place Vendôme, Notre Dame de Paris, Saint Germain Church Great Organ, and Palais of Versailles. FranceTelecom R&D recently built a research team dealing with "haptic" interfaces and their uses within the new telecommunications services such as "haptic" remote ultra-sound examination, haptic network games or virtual reality objects manipulation. Universitat Pompeu Fabra, SpainMultimedia Applications Group (DMAG) - Technology DepartmentThe Universitat Pompeu Fabra (UPF) is an only 12 years old University, but with already high reputation in lecturing and research. The UPF has 4 years old technical studies and research in IT, but with experienced researchers coming from other Universities or industry. The DMAG is a research group in the Technology Department, with members having more than 15 years experience in the field, including many European and national research projects (Multimediator, TRADE, MobiHealth, etc.). The main Research skills of DMAG include:
The UPF role in the network will be related to its skills and knowledge in the area of metadata and security, that will be applied to new environments, such as real time distribution of audiovisual content. The experience in electronic commerce of multimedia content will be also key to contribute to the network objectives. Another reason for the UPF to participate in the network is the existence of previous collaboration with some of the partners. This NoE will be a perfect opportunity to increase and formalise these relationships and to establish new ones with other partners. Finally, the experience of some of the DMAG members in international standardisation will be offered to the Consortium to cooperate in existing and future work items in the context of networked audiovisual systems, not only in the MPEG standardisation area, but also in others at different levels (CEN, ISO/IEC, …). Politecnico di Milano, ItalyDipartimento di Elettronica e InformazioneWith more than 39,000 engineering students enrolled, the Politecnico di Milano is the largest Technical University in Italy, as it graduates about 20% of the Italian engineers. The Dipartimento di Elettronica e Informazione (DEI) is among the largest ones of PdM, as it has about 160 professors. The main research activities of DEI are in the fields of telecommunications, multimedia signal processing, computer science, automation, robotics and artificial intelligence. The Image and Sound Processing Group of DEI-PdM, involved in the VISNET proposal, has a long-standing experience in the fields of image analysis, computer vision, audio-video processing, multimedia coding and compression. In particular, ISPG's expertise concerns geometric 3D reconstruction from multiple images acquired with calibrated and uncalibrated systems; motion analysis from image sequences; video coding (MPEG4, H264, etc.); sound synthesis based on physical modelling; fast geometry-based algorithms for sound spatialization, video analysis for tele-surveillance applications. ISPG-PdM is currently involved or has been involved in a number of national or international projects. In particular, among the EC-sponsored project involvements that are close in spirit to this proposal, it is worth mentioning:
The expertise and research interests of ISPG-PdM are perfectly matched with the areas covered by the VISNET Network of Excellence. The contribution that ISPG will give to the VISNET Consortium is in, particular, in the areas of sound generation and rendering, scalable video coding based on spatio-temporal wavelets and face detection and recognition. In all such areas, ISPG-PdM has already played an active role within several EC projects, and it foresees a tremendous potential benefit from the cooperation of the other partners of this Network of Excellence. Such benefit will be in terms of shared knowledge, tool development and exchange, joint software library development, etc. Warsaw University of Technology, PolandFaculty of Electronics and Information TechnologyWarsaw University of Technology is the largest technical university in Poland, top ranked for teaching and research in almost every field of technology. Faculty of Electronics and Information Technology, the largest faculty of Warsaw University of Technology, carry on educational, scientific and research activities covering almost the whole spectrum of relevant fields - from microelectronics and optoelectronics, biomedical engineering, through instrumentation, measurement, control and robotics, to computer engineering and networks, multimedia techniques and telecommunication. The university is represented in VISNET by Multimedia Group (IPW) headed by Professor Skarbek known for his contributions in the fields of image compression, analysis, indexing, and recognition. He is author of more than 100 scientific papers, three monographs, and chapters in Springer books. Significant part of his works had been developed while his stay abroad, in Japan and England. He chaired the 9-th international conference on Computer Analysis and Patterns CAIP’2003 and is the member of several conference and journal scientific boards. His group actively contributes to MPEG-7 visual subgroup for which it received special thanks in resolutions of Pattaya and Klagenfurt meetings in 2002. IPW group took part in many successful research projects for instance: CRIT2 – EU project on compression of medical images; multimedia monitoring and surveillance; multimedia home platform; H26L decoder. Beside Professor Skarbek, two Associate Professors, one Assistant Professor, and four doctoral level students will take part in integration projects of the VISNET consortium. IPW will coordinate WP1 on content creation and actively contribute to WP2 and WP4. Similarly to all VISNET partners, IPW will also make a contribution to the integration activities of WP0 and the dissemination activities of WP6. Instituto de Engenharia de Sistemas e Computadores do Porto, Portugal
INESC Porto - Instituto de Engenharia de Sistemas e Computadores do Porto is a
private, non-profit distributing association, whose partners are INESC, the
University of Porto and its School of Engineering. INESC Porto acts as an
interface between the academic world and the Information Technology and
Electronics sector, carrying out scientific research and development as well as
technology transfer and advanced professional training. The activities of INESC
Porto are carried out by five R&D Units (Telecommunications and Multimedia,
Optoelectronics and Electronic Systems, Power Systems, Manufacturing Systems
Engineering, Information and Communication Systems), under contract with
industry and services and in research projects funded by National Agencies and
EC R&D programmes.
|
last updated: 2008-05-19