Regional coverage is one of the most important tasks of public service broadcasting and an area that no other broadcaster covers to the same extent. In an era of increasing media consumption, this regionalisation requires resources that reduce and automate the effort involved. This begins at the broadcaster with the processing of documents or the technical preparation of content, but is also to be further advanced in post-production. rbb works closely with government institutions, service providers and research organisations to develop innovative solutions for all areas of the broadcaster. The focus is on supporting editorial teams to work more efficiently.

rbb used DeepVA as part of its process optimisation strategy in post-production, editing and archiving. Among other things, the test focused on landmark recognition to assist archive staff with keywording. DeepVA offered a solution that used AI to speed up, simplify and even partially automate the analysis of metadata. This enabled editors to access the material they needed more quickly and efficiently, and reduced repetitive tasks.

With just a few training images, the AI can be taught regional buildings or points of interest; this few-shot learning makes it very easy to get started with AI technology, and can be done entirely in the archive, without additional staff or external service providers, changing the job description of archive staff from archivist to AI manager – an added incentive as other departments within the broadcaster can also use the AI expertise and training data built up in the archive.r