Publication: Algoritmo de detección de manipulaciones inter-fotogramas en vídeos digitales
Loading...
Official URL
Full text at PDC
Publication Date
2021-09
Authors
Advisors (or tutors)
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Citations
Exportar
Abstract
Actualmente los dispositivos móviles están muy presentes en nuestra vida cotidiana.Cada día se captura una cantidad importante de imágenes y vídeos con este formato.A causa de esto, también aumenta el uso de herramientas de edición que permiten lamanipulación de la información capturada con este medio. Por lo que cobra especialinterés en el análisis forense el procesamiento de los datos provenientes de esta fuente.En este trabajo se desarrollan dos algoritmos que permiten identificar si existe la dichamanipulación en los vídeos provenientes de los móviles. El tipo de manipulación analizadoes el interfotograma. Se detecta tanto el barajado de fotogramas como la inserción oclonación de éstos. Se propone un algoritmo basado en una red neuronal convolucional para detectar la doble compresión y el algoritmo principal que detecta las manipulaciones en el vídeo.
Nowadays, mobile devices are very present in our daily lives. Every day a significant amount of images and videos are captured with this format. Because of this, the use of editing tools that allow the manipulation of the information captured with this medium is also increasing. Therefore, the processing of data from this source is of special interest in forensic analysis. Nowadays, mobile devices are very present in our daily lives. Every day a significant amount of images and videos are captured with this format. Because of this, the use of editing tools that allow the anipulation of the information captured with this medium is also increasing. Therefore, the processing of data from this source is of special interest in forensic analysis. Two algorithms are developed in this work that allow us to identify whether there is such manipulation in the videos from cell phones. The type of manipulation analyzed is the interframe. Both frame shuffling and frame insertion or cloning are detected. An algorithm based on a convolutional neural network is proposed to detect the double compression and the main algorithm that detects the manipulations in the video.
Nowadays, mobile devices are very present in our daily lives. Every day a significant amount of images and videos are captured with this format. Because of this, the use of editing tools that allow the manipulation of the information captured with this medium is also increasing. Therefore, the processing of data from this source is of special interest in forensic analysis. Nowadays, mobile devices are very present in our daily lives. Every day a significant amount of images and videos are captured with this format. Because of this, the use of editing tools that allow the anipulation of the information captured with this medium is also increasing. Therefore, the processing of data from this source is of special interest in forensic analysis. Two algorithms are developed in this work that allow us to identify whether there is such manipulation in the videos from cell phones. The type of manipulation analyzed is the interframe. Both frame shuffling and frame insertion or cloning are detected. An algorithm based on a convolutional neural network is proposed to detect the double compression and the main algorithm that detects the manipulations in the video.
Description
Trabajo de Fin de Grado en Ingeniería Informática, Facultad de Informática UCM, Departamento de Ingeniería del Software e Inteligencia Artificial, Curso 2020/2021.