We are pleased to announce an extended version of the LinkedTV tv2avd service (a.k.a. “Aggregated multimodal media analysis service”). This new version integrates software from our partner CERTH for video scene segmentation analysis which is applicable in any type of content, contrary to the developed chapter segmentation techniques which are strictly adapted to the analysis requirements of the LinkedTV scenarios (since they were based on algorithms that rely on the detection/re-detection of specific visual cues that appear in the AVROTROS and RBB content only).
In other words, the updated/extended version of the service contains the tools for providing a higher level temporal segmentation (than shots) of any video other than the material used in the project scenarios, since now the LinkedTV project is finished and new clients may wish to analyze other types of content (e.g. for demonstration purposes).
See a video demo explaining the existing functionalities.