Published March 31, 2021
| Version v1
Dataset
Open
MeMAD Knowledge Graph
Description
The MeMAD Knowledge Graph implements the EBU Core ontology and contains structured descriptions of more than 90k Radio and TV programs as well as more than 100k parts (segments) from 100 channels made available by INA (France) and Yle (Finland). This represents more than 64k hours of content in French, Finnish and Swedish.
The dataset contains:
- the MeMAD ontology that extends the EBU Core ontology
- the MeMAD controlled vocabularies represented in SKOS that interlink the programs genres, themes and the person roles using and extending the referenced EBU controlled vocabularies
- a number of graphs encapsulating the legacy metadata coming from the INA Legal Deposit, the INA Professional Archive and the Yle archive
- a graph encapsulating the automatic multimodal content analysis performed on some of those programs: this includes results from automatic speech recognition (ASR), extracting and disambiguating named entities from the ASR, face recognition from the video frames, machine translation of the ASR and visual deep captions generation
Files
Files
(272.1 MB)
Name | Size | Download all |
---|---|---|
md5:ac7c3ea97fc947ce8e15be08dfd9f326
|
264.8 MB | Download |
md5:bcb8dcf7c4de773b34133c7317b14aa5
|
7.3 MB | Download |
md5:ad258eb85bc21d742548e09d79904d7d
|
4.5 kB | Download |
md5:0c7271c4fdf53f3a32043b0c8526cf4e
|
45.6 kB | Download |