Publication: Identificación de la fuente en ficheros multimedia de redes sociales
Loading...
Official URL
Full text at PDC
Publication Date
2016
Authors
Advisors (or tutors)
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Citations
Exportar
Abstract
El impacto en la vida cotidiana que tiene el uso de dispositivos móviles para capturar imágenes digitales alimentó un creciente interés en compartir fotos y vídeos a través de las redes sociales. Desde un punto de vista forense, los usuarios de Internet, y en especial de estas redes sociales, dejan continuamente rastros visuales de su presencia en Internet, ofreciendo una valiosa información útil para las investigaciones.
Las herramientas de análisis forense de imágenes digitales permiten recolectar evidencias de estas imágenes. Una de estas técnicas utiliza el patrón del ruido del sensor (SPN) que actúa como una huella digital en la imagen del sensor de la cámara que la generó.
En este trabajo se propone utilizar el patrón PRNU, combinado con técnicas de agrupamiento, para identificar al propietario de un dispositivo móvil, analizando las fotos que un usuario de redes sociales sube desde el mismo dispositivo. Relacionar una imagen con el propietario de un dispositivo puede ser muy útil en una variedad de casos forenses: ciberbullying, abuso infantil en línea, búsqueda de cámaras robadas, etc.
La evaluación de la propuesta presentada se realizó con fotos subidas a 3 redes sociales por usuarios creados para este fin. Los resultados obtenidos presentan una tasa de acierto elevada, y demuestran el beneficio de usarla como herramienta forense.
The impact on everyday life that has the use of mobile devices to capture digital images fueled a growing interest in sharing images and videos through social networks. From a forensic point of view, Internet users, and especially of these social networks, continuously leave visual traces of their presence on the Internet, providing valuable information useful for investigations. Forensic analysis tools allow digital imaging collect evidence of these images. One of these techniques uses the pattern sensor noise (SPN) that acts as a fingerprint image of the camera sensor that generated it. This paper proposes to use the PRNU pattern, combined with clustering techniques to identify the owner of a mobile device, analyzing the photos that a user of social networks uploads from the same device. Relate an image with the device owner can be very useful in a variety of forensic cases: cyberbullying, online child abuse, search for stolen cameras, etc. The evaluation of the proposal was made with pictures uploaded by users (created for this purpose) to 3 social networks. The results obtained show a high rate of success, and demonstrate the benefit of using it as a forensic tool.
The impact on everyday life that has the use of mobile devices to capture digital images fueled a growing interest in sharing images and videos through social networks. From a forensic point of view, Internet users, and especially of these social networks, continuously leave visual traces of their presence on the Internet, providing valuable information useful for investigations. Forensic analysis tools allow digital imaging collect evidence of these images. One of these techniques uses the pattern sensor noise (SPN) that acts as a fingerprint image of the camera sensor that generated it. This paper proposes to use the PRNU pattern, combined with clustering techniques to identify the owner of a mobile device, analyzing the photos that a user of social networks uploads from the same device. Relate an image with the device owner can be very useful in a variety of forensic cases: cyberbullying, online child abuse, search for stolen cameras, etc. The evaluation of the proposal was made with pictures uploaded by users (created for this purpose) to 3 social networks. The results obtained show a high rate of success, and demonstrate the benefit of using it as a forensic tool.
Description
Trabajo de Fin de Grado en Ingeniería Informática (Universidad Complutense, Facultad de Informática, curso 2015/2016)