Publication: Técnica de detección de copia y empalme en imágenes y vídeos basada en redes neuronales
Loading...
Official URL
Full text at PDC
Publication Date
2020-09
Authors
Advisors (or tutors)
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Citations
Exportar
Abstract
Los vídeos se han convertido en un tipo de contenido ampliamente usado por la sociedad en todo tipo de ámbitos. El desarrollo tecnológico de los dispositivos móviles y la popularización de las plataformas web y redes sociales, han favorecido a que gran parte de los vídeos digitales provengan de dispositivos móviles. Además, el avance tecnológico nos solo nos permite grabar y difundir su contenido, sino también la edición y manipulación intencionada de los mismos.
Las técnicas de análisis forense de imágenes y vídeos de dispositivos móviles cobran pues, especial importancia. En este Trabajo Fin de Grado se propone un algoritmo que utiliza técnicas de inteligencia artificial para detectar la inserción y borrado de objetos en imágenes y vídeos. La técnica combina la extracción de características del contenido del fichero multimedia con el uso de redes neuronales simétricas. Para evaluar la eficiencia de la técnica propuesta se realizaron experimentos utilizando con conjunto de datos públicos.
Videos have become a widely used media file by society to spread news in all kind of areas. The technological development of mobile devices and the popularization of web platforms and social networks have promoted that a large part of digital videos are recorded using hand from mobile devices. In addition, technological progress allow you its distribution and also their manipulation. Therefore forensic analysis of images and videos of mobile devices gains special importance. This research work proposses an artificial intelligence algorithm based on the techniques used to detect the insertion and deletion of objects in images and videos files. The technique combines of the extraction of characteristics from the multimedia file with the use of symmetric neural networks. In order to evaluate the efficiency of the proposed technique, experiments were carried out using a public data set.
Videos have become a widely used media file by society to spread news in all kind of areas. The technological development of mobile devices and the popularization of web platforms and social networks have promoted that a large part of digital videos are recorded using hand from mobile devices. In addition, technological progress allow you its distribution and also their manipulation. Therefore forensic analysis of images and videos of mobile devices gains special importance. This research work proposses an artificial intelligence algorithm based on the techniques used to detect the insertion and deletion of objects in images and videos files. The technique combines of the extraction of characteristics from the multimedia file with the use of symmetric neural networks. In order to evaluate the efficiency of the proposed technique, experiments were carried out using a public data set.
Description
Trabajo de Fin de Grado en Ingeniería Informática, Facultad de Informática UCM, Departamento de Ingeniería del Software e Inteligencia Artificial, Curso 2019/2020.