Browsing by Subject "Xarxes neuronals (Informàtica)"
Now showing items 1-20 of 947
-
2D to 3D body pose estimation for sign language with Deep Learning
(Universitat Politècnica de Catalunya, 2020-06)
Bachelor thesis
Open AccessThis project aims at leveraging the challenge of using 3D poses for Sign Language translation or animation by transforming 2D pose datasets into 3D ones. The goal is, using a 3D dataset of American Sign Language, to train ... -
2D–3D geometric fusion network using multi-neighbourhood graph convolution for RGB-D indoor scene classification
(Elsevier, 2021-12)
Article
Open AccessMulti-modal fusion has been proved to help enhance the performance of scene classification tasks. This paper presents a 2D-3D Fusion stage that combines 3D Geometric Features with 2D Texture Features obtained by 2D ... -
3D car detection with LIDAR
(Universitat Politècnica de Catalunya, 2019-06)
Bachelor thesis
Open Access
Covenantee: Ficosa ElectronicsEn esta tesis se propone entrenar una red neuronal para la detección de coches, utilizando sólo puntos de un LIDAR de 32 canales de la base de datos de NuScenes. El objetivo es conseguir el reconocimiento de coches frontales ... -
3D convolutional neural networks for brain tumor segmentation
(2016)
Conference lecture
Restricted access - publisher's policyThis paper presents our work on applying 3D Convolutional Networks for brain tumor segmentation for the BRATS challenge. We are currently experimenting with different 3D fully convolutional architectures. We present ... -
3D convolutional neural networks for brain tumor segmentation: a comparison of multi-resolution architectures
(Springer, 2017-04-24)
Part of book or chapter of book
Open Access -
3D Face recognition with stereo automotive cameras
(Universitat Politècnica de Catalunya, 2019-06)
Bachelor thesis
Restricted access - confidentiality agreementThe automotive sector is experiencing a rapid change that will influence the way in which we will use cars. One of the main concerns of this new wave of automotive research regards the security of future vehicles, being ... -
3D point cloud correspondences using deep learning
(Universitat Politècnica de Catalunya, 2018-05)
Bachelor thesis
Open AccessThe main goal of the project is to find correspondences between points in two 3D point clouds using deep learning. A deep learning network is trained to select key points and find correspondences between two point clouds ... -
3D point cloud correspondences using graph neural networks
(Universitat Politècnica de Catalunya, 2018-10-17)
Bachelor thesis
Open AccessThe purpose of this project is the study of neural networks, their training and application together with the creation of a database appropriate to the system. Specically, the 3D convolutional networks applied to a database ... -
3D pose estimation using convolutional neural networks
(Universitat Politècnica de Catalunya, 2015-10)
Master thesis
Open AccessThe present Master Thesis describes a new Pose Estimation method based on Convolutional Neural Networks (CNN). This method divides the three-dimensional space in several regions and, given an input image, returns the ... -
A benchmark for graph neural networks for computer network modeling
(Universitat Politècnica de Catalunya, 2019-06-27)
Master thesis
Open AccessToday, network operators still lack functional network models able to make accurate predictions of end-to-end Key Performance Indicators (e.g., delay).This thesis introduces the benchmark for computer network modeling using ... -
A BF16 FMA is all you need for DNN training
(Institute of Electrical and Electronics Engineers (IEEE), 2022-07-01)
Article
Open AccessFused Multiply-Add (FMA) functional units constitute a fundamental hardware component to train Deep Neural Networks (DNNs). Its silicon area grows quadratically with the mantissa bit count of the computer number format, ... -
A chunking mechanism in a neural system for the parallel processing of a propositional production rules
(Universitat Politècnica de Catalunya. Secció de Matemàtiques i Informàtica, 1995)
Article
Open AccessThe problem of extracting more compact rules from a rule-based knowledge base is approached by means of a chunking mechanism implemented via a neural system. Taking advantage of the parallel processing potentialities of ... -
A closer look at referring expressions for video object segmentation
(2023-01)
Article
Open AccessThe task of Language-guided Video Object Segmentation (LVOS) aims at generating binary masks for an object referred by a linguistic expression. When this expression unambiguously describes an object in the scene, it is ... -
A communication infrastructure for emulating large-scale neural networks models
(Springer, 2012)
Conference report
Restricted access - publisher's policyThis paper presents the SEPELYNS architecture that permits to in- terconnect multiple spiking neurons focused on hardware implementations. SEPELYNS can connect millions of neur ons with thousands of synapses per neuron ... -
A comparative study of different gradient approximations for Restricted Boltzmann Machines
(Universitat Politècnica de Catalunya, 2023-06-29)
Master thesis
Open Access
Covenantee: Universitat Rovira i Virgili. Escola Tècnica Superior d'Enginyeria / Universitat de Barcelona. Facultat de Matemàtiques i InformàticaThis project consists of the theoretical study of Restricted Boltzmann Machines(RBMs) and focuses on the gradient approximations of RBMs. RBMs suffer from the dilemma of accurate learning with the exact gradient. Based on ... -
A Comprehensive survey on deep future frame video prediction
(Universitat Politècnica de Catalunya, 2018-01-25)
Master thesis
Open Access
Covenantee: Universitat de BarcelonaEl present projecte planteja l'estudi comprensiu i extens per a la tasca de predicció de fotogrames donada una seqüència de vídeo. Mitjançant l'anàlisi de l'estat de l'art en generació d'imatges, xarxes convolucionals i ... -
A cross-layer review of deep learning frameworks to ease their optimization and reuse
(Institute of Electrical and Electronics Engineers (IEEE), 2020)
Conference report
Open AccessMachine learning and especially Deep Learning (DL) approaches are at the heart of many domains, from computer vision and speech processing to predicting trajectories in autonomous driving and data science. Those approaches ... -
A Deep learning method for optimal stopping problems
(Universitat Politècnica de Catalunya, 2023-07)
Bachelor thesis
Open AccessL’objectiu principal d’aquest treball és complementar els fonaments teòrics i implementar el model de Deep Learning presentat a l’article anomenat “Deep Optimal Stopping”, de Becker et al, publicat al Journal of Machine ... -
A digital memristor emulator for FPGA-based artificial neural networks
(Institute of Electrical and Electronics Engineers (IEEE), 2016)
Conference report
Restricted access - publisher's policyFPGAs are reconfigurable electronic platforms, well-suited to implement complex artificial neural networks (ANNs). To this end, the compact hardware (HW) implementation of artificial synapses is an important step to obtain ... -
A framework for artificial data generation based on anatomical differences for electroencephalography-based brain-computer interfaces
(Universitat Politècnica de Catalunya, 2019-07-01)
Master thesis
Open Access
Covenantee: Technische Universität BerlinOne of the major limitations of brain-computer interfaces (BCI) is the need for a long and tedious calibration period in order for a subject to become proficient with the system. A principal challenge in training a BCI ...
