Publication: Clasificación de escenas en contenido audiovisual
Loading...
Identifiers
Publication date
2012-11
Defense date
2012-11-19
Authors
Advisors
Tutors
Journal Title
Journal ISSN
Volume Title
Publisher
Impact
Export
Abstract
En este proyecto se han diseñado algoritmos para la clasificación de escenas mediante categorías genéricas tales como: costa, calle, montaña, cielo, oficina, casa, etc. El objetivo de los algoritmos desarrollados es ser integrado en un sistema de anotación automático de contenidos visuales. Para ello, se realiza un estudio comparativo y sistemático de dos alternativas del estado del arte: el método de extracción de esencia de imágenes, llamado gist descriptor, y el método bag of words para la extracción de características. Para la clasificación se utiliza, además, una máquina de vector soporte. La motivación del proyecto es, en primer lugar, desarrollar un sistema que pueda ser capaz de clasificar una gran base de datos de imágenes según el tipo de escena visual, y, en segundo lugar, que el estudiante conozca y sepa utilizar diferentes métodos de clasificación de escenas audiovisuales, pueda evaluarlas y tomar conclusiones. ___________________________________________________________________________________________________________________________
In this project I had designed algorithms for scene classification using generic categories such as: coast road, mountain, sky, office, home, etc.. The developed algorithms have been integrated into a system for automatic annotation of visual content. For this purpose, I have performed a comparative and systematic study of two state of the art alternatives: the gist extraction method of images, called gist descriptor, and the bag of words method for feature extraction. For the classification, additionaly, I use a support vector machine. This project aims to, on the one hand, develop a system that will be able to classify a large database of images according to the type of audiovisual scene, and, on the other, that the student learns and knows how to use different methods of classification for visual scenes, as well as he can evaluate and draw his own conclusions from the results.
In this project I had designed algorithms for scene classification using generic categories such as: coast road, mountain, sky, office, home, etc.. The developed algorithms have been integrated into a system for automatic annotation of visual content. For this purpose, I have performed a comparative and systematic study of two state of the art alternatives: the gist extraction method of images, called gist descriptor, and the bag of words method for feature extraction. For the classification, additionaly, I use a support vector machine. This project aims to, on the one hand, develop a system that will be able to classify a large database of images according to the type of audiovisual scene, and, on the other, that the student learns and knows how to use different methods of classification for visual scenes, as well as he can evaluate and draw his own conclusions from the results.
Description
Keywords
Clasificación de escenas, Algoritmos, Gestión de contenidos, Bases de datos de imágenes