IDMT FRAUNHOFER INTRODUCES A/V TOOLBOX
In order to improve search and recommendation capabilities for the vast and ever-increasing amount of content within systems, it helps to have cost-effective technologies for automatic extraction of metadata from all that raw media. A typical use case can be searching video segments within a database where a specific speaker talks about a specific topic (keyword). Being one of the partners in the EU Project ‘MICO’, Fraunhofer IDMT was responsible for AV extractor development, integration and orchestration, and for the development of cross-media recommenders.
One of the core challenges in this domain is the need to shift from using individual standalone extractors for annotation to a multimodal and context-aware analysis workflow, which can provide substantially improved search functionalities and results.
To realise this potential, Fraunhofer IDMT developed solutions for:
- Cost-effective integration of heterogeneous extractors
- Flexible extractor workflow orchestration
- An extensible, common metadata model
- Availability of an expressive, media-aware query language
- Recommender systems using both annotations and collaborative filtering
With the newest AV analysing tools processes for broadcasting, distributing, and archiving digital content can be optimised. The newest AV analysing tools processes for broadcasting optimises distribution and archiving of digital content. The toolbox comprises various components for quality control, music and speech detection, video segment matching as well as video annotation tools for actor recognition, temporal video segmentation, and motion analysis. In addition, the audio forensic component allows identifying devices and microphones used for recordings, and the quick validation of digital content in terms of whether the content has been edited or encoded at some point in time.
AV Toolbox: www.idmt.fraunhofer.de/en/institute/projects_products/a_d/av_analyzing_toolbox.html