Publication: Identificación de la fuente de adquisición de ficheros multimedia de dispositivos móviles mediante Deep Learning
Loading...
Official URL
Full text at PDC
Publication Date
2019
Authors
Advisors (or tutors)
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Citations
Exportar
Abstract
Actualmente, la sociedad vive rodeada de contenido multimedia como son las imágenes y los vídeos. La presencia de dispositivos electrónicos capaces de realizar fotografías o grabar videos es una realidad en nuestra vida cotidiana, y su número aumenta con el paso del tiempo. La gran mayoría de la sociedad lleva un móvil en el bolsillo y hace uso de él para realizar fotos o vídeos. Ligado a ello, con los años han ido apareciendo técnicas de falsificación y manipulación de contenidos multimedia que dificultan saber si ese contenido es auténtico o no y de dónde procede, lo que hace que las técnicas de análisis forense sean una necesidad actual. En este trabajo se propone una red neuronal convolucional capaz de identificar la fuente de adquisición de vídeos grabados con un dispositivo móvil.
Los resultados obtenidos de los experimentos realizados en este trabajo demuestran la eficiencia de métodos propuestos. Para la evaluación de los métodos propuestos se realizaron experimentos con un dataset público ampliamente utilizado en la literatura y un dataset generado.
Nowadays, society lives surrounded by multimedia content such as images and videos. The presence of electronic devices capable of taking photographs or recording videos is a reality in our daily lives, and their number increases with the passage of time. The vast majority of society carries a mobile phone in their pocket and makes use of it to take photos or videos. Associated to this, over the years have been appearing falsification techniques and manipulation of multimedia content that make it difficult to know if that content is authentic or not and where it comes from, which makes forensic analysis techniques a current necessity. In this work a convolutional neuronal network capable of identifying the source of acquisition of videos recorded with a mobile device is proposed. The results that have been obtained in the experiments of this work shown clearly the effectiveness of the proposed methods. For the evaluation of these proposed methods there have been carried out experiments using a public dataset, which has been amplified with the literature and a generated dataset.
Nowadays, society lives surrounded by multimedia content such as images and videos. The presence of electronic devices capable of taking photographs or recording videos is a reality in our daily lives, and their number increases with the passage of time. The vast majority of society carries a mobile phone in their pocket and makes use of it to take photos or videos. Associated to this, over the years have been appearing falsification techniques and manipulation of multimedia content that make it difficult to know if that content is authentic or not and where it comes from, which makes forensic analysis techniques a current necessity. In this work a convolutional neuronal network capable of identifying the source of acquisition of videos recorded with a mobile device is proposed. The results that have been obtained in the experiments of this work shown clearly the effectiveness of the proposed methods. For the evaluation of these proposed methods there have been carried out experiments using a public dataset, which has been amplified with the literature and a generated dataset.
Description
Trabajo de Fin de Grado en Doble Grado en Ingeniería Informática y Matemáticas, Facultad de Informática UCM, Departamento de Ingeniería del Software e Inteligencia Artificial, Curso 2018/2019.