Fraunhofer IDMT’s presentation will include various technologies for automatic metadata extraction, quality control, audio forensics, and privacy-aware data analysis and recommendation. This year’s focus will be on improved fingerprinting-based analysis technologies for both query-based matching and partial matching to support broadcasters and archives in detecting occurrences and relationships within content sets.

Increased robustness for query-based matching

Query-based matching can be used for broadcast monitoring, including ad detection to find occurrences of a given audio query within a stream or file. At IBC 2018, Fraunhofer IDMT shows a further-improved version of this technology: The software detects matches even if the sample file is noisy due to mobile capturing and air transmission or if the file is used in a mix and other signals are present within the stream.

Segment matching without queries

Other application domains require detection of partial matches within a set of items without queries, that means without prior knowledge about whether and which overlaps exist. Fraunhofer IDMT’s segment matching technology automatically detects reuse of material within production archives and helps to propagate metadata or to determine and clear rights for a production. Furthermore, it can identify content reuse within and among A/V streams to get a quick overview over a program structure as a basis for further analysis. For news search and production, it is useful to quickly detect material in and among archives which shares the same original footage and hence stems from the same event. And finally, segment matching technology helps to identify and remove partial duplicates within archives.

Visit us at the Fraunhofer booth at 8B80 to learn more about the aforementioned technologies and applications.