Virginia Bazán-Gil

Virginia Bazán-Gil

Auto-generated profile, based on publicly available information

Virginia Bazán-Gil

Corporación Radiotelevisión Española, 28223 Madrid, Spain

Research Focus & Metrics

Academic Career Length

8Years (2019-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
11 December 2019 in Applied Sciences
Verified
Journal Article
Open access
PDF file
ELEduardo Lleida
Eduardo Lleida
AOAlfonso Ortega
Alfonso Ortega
AMAntonio Miguel
Antonio Miguel
VBVirginia Bazán-Gil
Virginia Bazán-Gil
MPMª Carmen Pérez
Mª Carmen Pérez
APAlberto de Prada
Alberto de Prada
The IberSpeech-RTVE Challenge presented at IberSpeech 2018 is a new Albayzin evaluation series supported by the Spanish Thematic Network on Speech Technologies (Red Temática en Tecnologías del Habla (RTTH)). That series was focused on speech-to-text transcription, speaker diarization, and multimodal diarization of television programs. For this purpose, the Corporacion Radio Television Española (RTVE), the main public service broadcaster in Spain, and the RTVE Chair at the University of Zaragoza made more than 500 h of broadcast content and subtitles available for scientists. The dataset included about 20 programs of different kinds and topics produced and broadcast by RTVE between 2015 and 2018. The programs presented different challenges from the point of view of speech technologies such as: the diversity of Spanish accents, overlapping speech, spontaneous speech, acoustic variability, background noise, or specific vocabulary. This paper describes the database and the evaluation process and summarizes the results obtained.

26 September 2023 in Profesional de la información
Verified
Journal Article
Open access
PDF file
VBVirginia Bazán-Gil
Virginia Bazán-Gil
The aim of this paper is to present an international overview of the use of artificial intelligence in the context of media archives in broadcasters, preservation institutions and press agencies, through a comprehensive analysis of sources primarily focusing on case studies presented at international conferences and seminars, together with the results of the survey on the use of artificial intelligence conducted by FIAT/IFTA. Once the most commonly used technologies have been defined and we have identified the stages of the production workflow in which they are used, we will discuss the specific applications of these technologies in television archives, audiovisual heritage preservation organisations, press agencies and innovation projects where technology vendors and media companies collaborate. Finally, we will deal with the challenges related to the implementation of AI in media archives, the need for datasets in the development of language models, and the relevance of a sensible use of technology.

14 June 2021 in Profesional de la información
Verified
Journal Article
Closed access
PDF file
VBVirginia Bazán-Gil
Virginia Bazán-Gil
MZMANUEL GOMEZ ZOTANO
MANUEL GOMEZ ZOTANO
NMNoemí Marroyo-Núñez
Noemí Marroyo-Núñez
PSPaloma Sampedro-Canet
Paloma Sampedro-Canet
DDDavid De-Ignacio-Ledesma
David De-Ignacio-Ledesma
The results of a project on news segmentation at Radio Nacional de España (RNE) carried out by the RTVE Technological Innovation and Media Management areas is presented. The aim of this project is to apply artificial intelligence to automatically transcribe and cut the news items that make up a radio news program. The main goals of this project are to increase the accessibility of the content and to allow its reusability on various platforms and social media. The project was planned in two phases, covering system configuration and service delivery. The minimum quality criteria required were defined in advance, both for automatic voice transcription and for news segmentation. For the speech-to-text process, the highest word error rate (WER) allowed was 10%, while the precision rate for the news segmentation was 85%. System performance in both transcription and segmentation was considered to be sufficient, although a higher degree of accuracy in news cutting is expected in the coming months. The results show that, despite using these quite mature technologies, adjustment and learning processes and human intervention are still necessary. Resumen Se presentan los resultados del proyecto para la segmentación en noticias de los informativos de Radio Nacional de España (RNE) llevado a cabo por el Área de Innovación Tecnológica de Radio Televisión Española (RTVE) en colaboración con la Dirección de Medios de RNE. El objetivo de este proyecto es aplicar la inteligencia artificial para el cortado automático de las noticias que componen un informativo radiofónico, para su posterior difusión en la web de RTVE y en medios de comunicación social. El proyecto se planificó en dos fases: una primera de configuración y ajuste del sistema, y una segunda de prestación del servicio propiamente dicho. Los criterios de calidad mínimos exigibles se definieron previamente, tanto para la transcripción automática del habla a texto, para la que se estableció una tasa de error por palabra máxima (WER) del 10%, como para la segmentación de noticias, para la que se definió una tasa de precisión superior al 85%. El rendimiento del sistema tanto en la transcripción como en la segmentación se considera suficiente, si bien se espera alcanzar un mayor grado de precisión en el cortado de noticias en los próximos meses. Los resultados ponen de manifiesto que, a pesar de ser tecnologías bastante maduras, son necesarios procesos de ajuste y aprendizaje con la intervención humana.

25 July 2023 in Applied Sciences
Verified
Journal Article
Open access
PDF file
ELEduardo Lleida
Eduardo Lleida
Luis Javier Rodriguez-FuentesLuis Javier Rodriguez-Fuentes
Luis Javier Rodriguez-Fuentes
Javier TejedorJavier Tejedor
Javier Tejedor
AOAlfonso Ortega
Alfonso Ortega
AMAntonio Miguel
Antonio Miguel
HAHaritz Arzelus
Haritz Arzelus
Evaluation campaigns provide a common framework with which the progress of speech technologies can be effectively measured. The aim of this paper is to present a detailed overview of the IberSpeech-RTVE 2022 Challenges, which were organized as part of the IberSpeech 2022 conference under the ongoing series of Albayzin evaluation campaigns. In the 2022 edition, four challenges were launched: (1) speech-to-text transcription; (2) speaker diarization and identity assignment; (3) text and speech alignment; and (4) search on speech. Different databases that cover different domains (e.g., broadcast news, conference talks, parliament sessions) were released for those challenges. The submitted systems also cover a wide range of speech processing methods, which include hidden Markov model-based approaches, end-to-end neural network-based methods, hybrid approaches, etc. This paper describes the databases, the tasks and the performance metrics used in the four challenges. It also provides the most relevant features of the submitted systems and briefly presents and discusses the obtained results. Despite employing state-of-the-art technology, the relatively poor performance attained in some of the challenges reveals that there is still room for improvement. This encourages us to carry on with the Albayzin evaluation campaigns in the coming years.

19 November 2024 in Tábula
Verified
Journal Article
Closed access
PDF file
VBVirginia Bazán-Gil
Virginia Bazán-Gil
Este artículo explora la integración de inteligencia artificial (IA) en el Archivo de RTVE para generar automáticamente metadatos y mejorar la accesibilidad de contenidos audiovisuales. La IA se ha implementado para optimizar la catalogación y recuperación de colecciones filmadas, especialmente los fondos más antiguos de RTVE. Desde las primeras pruebas en 2017 hasta la implementación de servicios en 2021 y 2023, se han mejorado 16.000 horas de contenido mediante tecnologías avanzadas de IA, como Whisper de OpenAI y GPT-3.5. Se describe la arquitectura del sistema, el flujo de archivos y los procesos de validación, abarcando tareas como transcripción automática, identificación de hablantes, reconocimiento facial y de caracteres, detección de objetos y logos, y descripción de imágenes. Se analizan resultados y desafíos futuros, destacando la importancia de casos de uso específicos para RTVE. La IA ha mejorado la infraestructura tecnológica del archivo, optimizando el metadatado y recuperación de contenidos. Se subraya la importancia de una comunicación transparente, formación continua del personal y la supervisión humana de la IA. A largo plazo, la combinación de IA y estrategias puede mejorar la preservación y accesibilidad de contenidos históricos valiosos. This article explores the integration of artificial intelligence (AI) in the RTVE Archive for automatic metadata generation and improved accessibility of audiovisual content. AI has been implemented to optimize cataloging and content retrieval, particularly for RTVE’s oldest filmed collections. From initial tests in 2017 to the implementation of services in 2021 and 2023, 16,000 hours of content have been enhanced using advanced AI technologies like OpenAI’s Whisper and GPT-3.5. The system’s architecture, file workflow, and validation processes are described, covering tasks such as automatic transcription, speaker identification, facial and optical character recognition, object and logo detection, and natural language image description. Results and future challenges are analyzed, emphasizing the importance of specific use cases for RTVE. AI has improved the technological infrastructure of the archive, optimizing metadata and content retrieval. The importance of transparent communication, continuous staff training, and human supervision of AI is highlighted. In the long term, combining AI with strategic actions can enhance the preservation and accessibility of valuable historical content.