Enviaments recents

  • Deep learning for universal emotion recognition in still images 

    Rosa Ramos, Juan Luis (Universitat Politècnica de Catalunya, 2018-04)
    Projecte Final de Màster Oficial
    Accés obert
    Realitzat a/amb:  Computer Vision Center / Universitat de Barcelona / Universitat Rovira i Virgili / High Performance Computing UAB
    This work propose a methodology for still image facial expression. The proposed method contains a face detection and alignment module followed by a deep convolutional neural network (CNN) that outputs a seven emotions ...
  • Using peptidomics and machine learning techniques to predict mortality of patients with septic shock 

    Byrne, Helen (Universitat Politècnica de Catalunya, 2018-04-26)
    Projecte Final de Màster Oficial
    Accés obert
    Realitzat a/amb:  ShockOmics, Eurecat / Facultat de Matemàtiques UB / Escola Tècnica Superior d'Enginyeria URV
    The main objective of this thesis is to analyse peptidomics data and evaluate its use in the mortality prediction of patients in septic shock. This is the first time this peptidome data has been analysed in this depth. A ...
  • Semantic distances between medical entities 

    Olivares Alarcos, Alberto (Universitat Politècnica de Catalunya, 2018-04)
    Projecte Final de Màster Oficial
    Accés obert
    In this thesis, three different similarity measures between medical entities (drugs) have been implemented. Each of those measures have been computed over one or more dimensions of the drugs: textual, ...
  • Clinical assessment of shock through machine learning techniques 

    Aushev, Alex (Universitat Politècnica de Catalunya, 2018-04)
    Projecte Final de Màster Oficial
    Accés obert
    This work attempts to identify biomarkers that can be used as indicators for mortality prediction in patients with shock. It uses machine learning techniques to analyze the clinical data, impute missing values in it, select ...
  • New hybrid kernel architectures for deep learning 

    Mora de Checa, Daniel (Universitat Politècnica de Catalunya, 2018-04)
    Projecte Final de Màster Oficial
    Accés obert
    In this work we explore the possibilities of combining neural network architectures and kernel methods by introducing hybrid kernel blocks. We present hybrid architectures which can be trained as traditional neural networks ...
  • Abstractive text summarization with attention-based mechanism 

    Sanjabi, Nima (Universitat Politècnica de Catalunya, 2018-04)
    Projecte Final de Màster Oficial
    Accés obert
    In this work, we explore the evolution of Sequential Neural Models, and their use as a Summarizer System. Transformer is a recently proposed model with a high potential. We experiment and compare their result in abstractive ...
  • Adaptative case-based reasoning: maintenance and learning strategies 

    Nakhjiri, Nariman (Universitat Politècnica de Catalunya, 2018-04)
    Projecte Final de Màster Oficial
    Accés restringit per acord de confidencialitat
    Realitzat a/amb:  Universitat de Barcelona
    In this master thesis, two novel families of methods in the field of Case-Base Maintenance have been presented. Both of the proposed methods have successful results and one of them even surpass well-known state-of-the-art ...
  • Natural language processing and machine learning techniques to solve a breast cancer clinical trial ECOG-classification problem 

    Reynoso Aguirre, Pablo Eliseo (Universitat Politècnica de Catalunya, 2018-04-16)
    Projecte Final de Màster Oficial
    Accés obert
    Natural Language Processing and Machine Learning Framework to Predict Breast Cancer Participant's Profile in Clinical Trials based on ECOG/KPS Performance Status Scores using Structured Research Data from All over the world.
  • Bridge Structural Damage Segmentation Using Fully Convolutional Networks 

    Rubio Guillamón, Juanjo (Universitat Politècnica de Catalunya, 2018-04-16)
    Projecte Final de Màster Oficial
    Accés obert
    Realitzat a/amb:  National Institute of Informatics
    Fully Convolutional Networks prove to be suitable method for texture-based damage segmentation on infrastructure. A dataset has been collected to model the uncertainty in human inspection of bridges in the Japanese prefecture ...
  • Social relation recognition in egocentric photo-strems 

    Sanchez Aimar, Emanuel (Universitat Politècnica de Catalunya, 2018-04-17)
    Projecte Final de Màster Oficial
    Accés obert
    Realitzat a/amb:  Universitat de Barcelona / Universitat Rovira i Virgili
    Presentamos un método para la categorización automática de relaciones sociales en secuencias de imágenes egocéntricas. Por cada imagen, extraemos características visuales con redes convolucionales (atributos como edad, ...
  • Ethical reasoning for norm-based systems 

    Serramià Amorós, Marc (Universitat Politècnica de Catalunya, 2018-04-16)
    Projecte Final de Màster Oficial
    Accés restringit per acord de confidencialitat
    Realitzat a/amb:  Universitat de Barcelona / Universitat Rovira i Virgili
    Deciding norms to rule a society is a difficult task. This work focuses on choosing the 'right' norms by considering moral values and the society's partial preferences over these values. Ethics and other considerations, ...
  • Multiple object tracking using RNNs 

    Amin, Khizer (Universitat Politècnica de Catalunya, 2017-12)
    Projecte Final de Màster Oficial
    Accés obert
    Tracking multiple objects simultaneously from a single camera is popular in surveillance systems. The standard methodology to address the problem is decomposing it in the object appearance modeling, object detection/tracking ...

Mostra'n més