End of STREAMS, a platform that deploys connected AI services for transcription, translation and speech synthesis in multiple languages for the generation of multiple applications
The project, in collaboration with a consortium of seven companies and with Vicomtech as scientific-technological coordinator, has as its main objective the construction of a platform with multilingual and connectable AI services for transcription, translation, speech synthesis and audio description that allow the construction of various applications for batch processing and streaming of multimedia content.
24.04.2024
It is a reality that the generation and content consumption through the Internet has skyrocketed in recent years, especially influenced by the digital transformation, new consumption habits and the ongoing technological evolution.
New advances in Artificial Intelligence applied to the processing and creation of digital content have also helped to increase the possibilities for a better, more democratised and accessible content for all. In this way, technological developments are increasingly making available new solutions that automate the processes of generating accessible content, in some cases enabling access to it and generally improving the user experience.
It is in this context that STREAMS was born, a project in collaboration with seven companies: Haiko, Jarkatza, Mondragon Lingua T&C, Goiena Komunikazio Taldea, Mixer Servicios Audiovisuales, Noticias de Gipuzkoa and Onda Vasca; in which Vicomtech collaborates as scientific-technological coordinator.
The solution consists of a platform that can be deployed locally or in the cloud, based on Artificial Intelligence technologies, with connected transcription, translation and speech synthesis services in four languages: Spanish, Basque, French and English.
STREAMS is aimed at providing and enabling agile and dynamic information transmission and generation flows in different application scenarios, such as micro-batching (small batch processing of files) and streaming (real-time processing).
Vicomtech provides its expertise in speech and machine translation technologies for the development of the STREAMS platform.
In collaboration with the seven companies of the consortium led by HAIKO, Vicomtech works as the scientific-technological coordinator of STREAMS.
The centre is incorporating technologies such as enriched speech recognition, machine translation, speech synthesis and automatic alignment in all four languages into the project.
Thus, by connecting these compatible technological modules in different processing pipelines, the platform enables the generation of multiple applications such as automatic multilingual subtitling in real time, automatic audio description, the generation of high quality synthetic audio, or the creation of transcriptions and translations in different output formats, among others.
Thanks to Vicomtech's knowledge in language technologies, the centre provides high-value technological solutions that, in the case of the STREAMS project, enhaces the competitiveness of companies in the digital and audiovisual content sector by offering an innovative and differentiating solution for multiple application areas.
The project is part of the Hazitek programme promoted by the SPRI Group and the Basque Government's Department of Economic Development, Sustainability and Environment.
