preprocessing clasification ddbb features extraction
play

Preprocessing Clasification DDBB Features extraction - PowerPoint PPT Presentation

MULTIMEDIA DATA ANALISYS FOR EMOTION RECOGNITION D e v e l o p a m e t h o d o l o g y t h a t a l l o w s t o d e t e c t t h e e m o t i o n s o f t h e p e o p l e t h r o u g h m u l t i m e d i a d a t a , i n s p e c i a l , s


  1. MULTIMEDIA DATA ANALISYS FOR EMOTION RECOGNITION D e v e l o p a m e t h o d o l o g y t h a t a l l o w s t o d e t e c t t h e e m o t i o n s o f t h e p e o p l e t h r o u g h m u l t i m e d i a d a t a , i n s p e c i a l , s p e e c h d a t a , u s i n g D e e p N e u r o n a l N e t w o r k s . • FFT • Arquitecture • Mel scale • Op9miza9on • Log transforma9on • Ac9va9on func9on • Mul9window Preprocessing Clasification ☺ � � … DDBB Features extraction • Pretraining • Transforma9ons. • CNN • Concatenate DDBB • RLU/Sigmoid • Adding noise • Incep9on Fernando García Novo, fgnovo@gmail.com

  2. LMS and H-SATCOM Radio Channel Characterization at X and KU bands  Wideband and narrowband study  Spatial and polarization diversity techniques  3D ray tracing model  Vegetation scattering model -60 0 Direct ray Blade 1 - trailing edge -65 -10 Blade 1 - transversal edge Blade 1 - leading edge normalized power to direct ray(dB) Blade 2 Blade 3 received power [dBm] -20 Blade 4 -70 -30 -75 Thuja Line fit -40 d fs -80 d obs Free space -50 -85 0 10 20 30 40 50 60 70 80 90 100 0 1 1.5 2 2.5 time(ms) distance between receiver and tree center [m]

  3. Behavioural Modelling of Microwave Transistors for Wideband High Efficiency Power Amplifier Design UAV Device 6x75 μ m GaN HEMT. Load-pull contours at 3 GHz and P1dB. Input Output “ - ” Interpolated Extrapolated Meas. system PNA-X based set-up Author: Mª del Rocío Moure Fernández PHD PROGRAM ON INFORMATION AND COMMUNICATIONS TECHNOLOGY OF Advisor: Mónica Fernández Barciela THE UNIVERSITY OF VIGO

  4. Secure Signal Processing for Genomic Privacy Protection Mina Namazi Motivation • Genomic data reveals each person’s carrier status, Results health risks, drug response, inherited traits, Sensitive. • Genomic data are unique identifier of each person. • Novel protocol of susceptibility • Genomic data are not revocable. test by applying lattice based cryptography Thesis Objectives • Achieving higher efficiency , less round complexity , and more • E-Health applications privacy analysis privacy • Designing privacy preserving enhanced signal • Sketch of the protocol featuring processing methods for e-health applications. an “inherent” access control • Devising new information-theoretic metrics to through attributes enforcing the quantify the information leakage on genomic patient’s access policy data. • Dynamic Privacy-Preserving Genomic Susceptibility Testing.

  5. SIGNAL PROCESSING FOR ANONYMOUS COMMUNICATIONS Simon Oya , Carmela Troncoso, and Fernando Pérez-González simonoya@gts.uvigo.es ctroncoso@gradiant.org fperez@gts.uvigo.es MOTIVATION OF THE WORK THESIS OBJECTIVES Apply signal processing tools ? to anonymous communications. Alice Analyze Develop Oncologist Optimize Privacy in the communications: Protect data (cryptography) How? In the poster :) Protect meta-data!!!

  6. Contribution to development of telematic services for data analysis in technology area. Application to eHealth field. Author: Mateo Ramos Merino Thesis Advisors: Juan M. Santos Gago, Luis M. Álvarez Sabucedo Department of Telematic Engineering , University of Vigo In some domains, such as eHealth, process monitoring is very important. It is mandatory to control, check and verify the implementation of workflows as designed in actual scenarios. Desired behavior Workflow Process model OBJECTIVES Improve expressiveness of Real behavior current modeling languages. Collect traces on this for analyses Enhance results of conformance checking techniques using semantic information.

  7. CONTRIBUTION TO THE VIRTUALIZATION AND ROUTING MECHANISMS FOR AUTONOMOUS DRIVING APPLICATIONS IN MOBILE AND VEHICULAR AD-HOC NETWORKS Author: José Víctor Saiáns-Vázquez Thesis Advisors: Martín López-Nores and Yolanda Blanco-Fernández MOTIVATION VANET’s problem → Solution High mobility Virtualization ► VANETs are expected to become an extension of the wired Internet , providing innovative communication and information services to the drivers and passengers. ► Virtualization is a good way to tackle the problem of the high mobility: ► C luster-based approach to handle communications. ► Abstraction of fixed geographical regions served by virtual nodes as a mean to tackle the mobility of the real nodes. ► The virtualization mechanisms can be polished to work efficiently in a wide variety of scenarios . ► Autonomous driving is one of the most promising application of the VANETs. THESIS OBJECTIVES

  8. APPLICATION OF LEARNING ANALYTICS TECHNIQUES ON BLENDED LEARNING ENVIRONMENTS FOR UNIVERSITY STUDENTS Sheila Lucero Sánchez López Supervised by: Rebeca P. Díaz Redondo, Ana Fernández Vilas Affiliation: ICLab Information & Computing Lab, Department of Telematics Engineering (University of Vigo) º The emergence of E-Learning platforms is innovating the learning process, the platforms store in its databases information about the actions of teachers and students, thus generating large volumes of data but has not been able to measure the degree of learning acquisition. Therefore, the main objective of this thesis is the research and application of learning analytics techniques for prediction and prevention of failure and dropout of students at university level under the blended-learning training model. We propose: Blended Learning Analytics enviroments “The measurement, collection, analysis and reporting of data about learners and their contexts, for purposes of understanding and optimizing learning and the environments in which it occurs ”. LAK 2011

  9. UNDERWATER NOISE MAPPING METHODOLOGIES FOR SHALLOW WATERS Author: David Santos-Domínguez | Thesis advisor: Soledad Torres-Guijarro | AffiliaIon: Sonitum (TSC, Universidad de Vigo) O1 Underwater noise measurement methodologies. O2 PropagaIon Losses calculaIon using both experimental and analyIc models. O3 Study of underwater noise predicIon soWware. O4 ClassificaIon of the different noise sources available in Ría de Vigo. O5 Noise map construcIon methodologies. O6 Construc1on of an underwater noise map of Ría de Vigo. Ría de Vigo underwater noise map recreaIon

  10. Recommendation of Tourism Resources Supported by Crowdsourcing Profiling & Crowdsourcing Trust & Reputation Recommendations  State-of-the-art review  Crowdsourcing  Build the user profile data  Tourist behaviour analysis (Tourists)  Identification  Recommendation questionnaire of of patterns and trends tourism resources  Usage of Wikivoyage based on the current to design and validate user context and Trust & Reputation profile mechanisms Tourism Mobile Applications Done Next Planning Year Fátima Leal, University of Vigo, June 2016 – Workshop DOC-TIC

  11. ICT IN PROMOTING BUSINESS TRAINING PhD in ICT thesis advisor MARTÍN LLAMAS NISTAL author HELDER RODRIGO SOARES PINTO ARCHITECTURE AND TELEMATIC SERVICES WEB & MOBILE E-LEARNING APPLICATIONS 35 WEB & BUSINESS MOBILE TRADICIONAL TRAINING BUSINESS HOURS TRAINING FLEXIBLE HOURS EASY ACCESS SPEND MORE TIME LOWER PRICES LEARNING PACE FOR THE COURSES TRAVEL EXPENSES SET BY THE TRAINEE LOW IN PRODUCTIVITY REDUCED TRAVEL PERMANENT AVAIL ABILITY OF CONTENT HIGHER PRICES DIVERSIFICATION COURSES

  12. Workshop on Monitoring PhD Student Progress A EARLY DETECTION OF COGNITIVE IMPAIRMENT June 13, 2016 THROUGH GAMIFICATION, MACHINE LEARNING 0 TECHNIQUES AND ICT TOOLS MAIN GOAL: To design and develop a DIGITAL TEST to detect early cognitive impairment http://panoramix.gist.det.uvigo.es/ Author: Sonia Mª Valladares Rodríguez [soniavr@det.uvigo.es] Advisors: Luis Anido Rifón [Full Professor] J. Manuel Fernández Iglesias [Associate Professor]

Download Presentation
Download Policy: The content available on the website is offered to you 'AS IS' for your personal information and use only. It cannot be commercialized, licensed, or distributed on other websites without prior consent from the author. To download a presentation, simply click this link. If you encounter any difficulties during the download process, it's possible that the publisher has removed the file from their server.

Recommend


More recommend