Publication: Módulo de transcripcción VoIP - texto para una plataforma de interceptación legal de comunicaciones
Loading...
Identifiers
Publication date
2012
Defense date
2012-02-22
Authors
Advisors
Tutors
Journal Title
Journal ISSN
Volume Title
Publisher
Impact
Export
Abstract
En este proyecto se estudia, implementa y evalúa un módulo de transcripción VoIP-texto para una plataforma de interceptación legal de comunicaciones. Esta plataforma consiste en un software de análisis forense, llamado Xplico, que permite analizar el tráfico de red. Es software libre, y permite clasificar por categorías (tipo de contenido) la información capturada de la red. Ejemplos de estas categorías pueden ser páginas web, imágenes, correo, llamadas VoIP, intercambio de archivos, etc. El objetivo principal del proyecto es ahorrar tiempo y trabajo a los analistas de la Policía en la tarea de analizar conversaciones de voz sobre IP (VoIP – Voice over IP) llevadas a cabo por sospechosos, que estén siendo monitorizados por orden judicial. De modo que el proyecto consiste, a grandes rasgos, en la incorporación de un software de reconocimiento de voz open source (Sphinx-4) a la plataforma descrita anteriormente. Como se ha señalado, Xplico es capaz de interceptar flujos de datos VoIP, pero hasta ahora, no era posible procesar esos datos automáticamente. Gracias a la incorporación del software de Sphinx-4, el sistema es capaz de transcribir automáticamente las conversaciones interceptadas, proporcionando además, información detallada sobre los interlocutores que intervienen en la conversación. Además de lo citado anteriormente, se ofrece la posibilidad de entrenar el sistema con el objetivo de obtener mejores resultados (mayor precisión en la transcripción) en el futuro. ____________________________________________________________________________________________________________________
In the present project, VoIP-text transcription module for a legal interception of communications platform has been studied, implemented and evaluated. This platform is forensic analysis software, called Xplico, to analyze network traffic. It’s free software, and allows categorizing the information captured from the network. Examples of these categories may be web pages, images, mail, VoIP calls, file sharing, etc. The main goal of the project is to save time and effort to the police’s analysts in the task of analyzing voice conversations over IP (VoIP - Voice over IP) carried out by suspects who are being monitored by court order. So the project is, broadly speaking, the addition of an open source speech recognition software (Sphinx-4) to the platform described above. As noted, Xplico is able to intercept VoIP data streams, but so far, it was not possible to process these data automatically. With the addition of Sphinx-4 software, the system can automatically transcribe the conversations intercepted, providing also, details on the speakers involved in the conversation. Besides the above, it offers the possibility to train the system in order to obtain better results (increased accuracy in transcription) in the future.
In the present project, VoIP-text transcription module for a legal interception of communications platform has been studied, implemented and evaluated. This platform is forensic analysis software, called Xplico, to analyze network traffic. It’s free software, and allows categorizing the information captured from the network. Examples of these categories may be web pages, images, mail, VoIP calls, file sharing, etc. The main goal of the project is to save time and effort to the police’s analysts in the task of analyzing voice conversations over IP (VoIP - Voice over IP) carried out by suspects who are being monitored by court order. So the project is, broadly speaking, the addition of an open source speech recognition software (Sphinx-4) to the platform described above. As noted, Xplico is able to intercept VoIP data streams, but so far, it was not possible to process these data automatically. With the addition of Sphinx-4 software, the system can automatically transcribe the conversations intercepted, providing also, details on the speakers involved in the conversation. Besides the above, it offers the possibility to train the system in order to obtain better results (increased accuracy in transcription) in the future.
Description
Keywords
Protocolos de comunicación, Reconocimiento de voz