Synthetic 2D point clouds generator
Consultable a partir de
2024-10-01
Fecha
2019Director
Versión
Acceso embargado 5 años / 5 urteko bahitura
Tipo
Trabajo Fin de Máster/Master Amaierako Lana
Impacto
|
nodoi-noplumx
|
Resumen
Every day, while developing applications and products to solve a huge number of
today problems, data from real world is registered and consumed. The registration of this
kind of data is costly, and on its quality depends on the correct behaviour of the solutions
developed. Some of this data are 2-dimensional point clouds, for example spatial points
registered by sensors. In this project, we p ...
[++]
Every day, while developing applications and products to solve a huge number of
today problems, data from real world is registered and consumed. The registration of this
kind of data is costly, and on its quality depends on the correct behaviour of the solutions
developed. Some of this data are 2-dimensional point clouds, for example spatial points
registered by sensors. In this project, we present and investigate the use Generative
Adversarial Networks and Neural Style Transfer over 2-dimensional point clouds in order
to develop a tool to generate synthetic but realistic data based on real ones. We also
study the possibility of combining these two technologies to improve each other's
behaviour. [--]
Materias
Generative adversarial networks,
Neural style transfer,
2D point clouds,
Automotive industry,
Deep learning
Titulación
Máster Universitario en Ingeniería de Telecomunicación por la Universidad Pública de Navarra /
Nafarroako Unibertsitate Publikoko Unibertsitate Masterra Telekomunikazio Ingeniaritzan