Indexing & Search

Veridict Large Vocabulary Recognition System enables automatic transcription of audio and video material to indexed, searchable text. On live material, e.g. video feeds and news broadcasts, text transcriptions are obtained time-synchronously, in near real-time (see demonstration video below). In offline mode, where large databases are to be transcribed to text, system runtime is tens or hundreds times faster than real-time. This enables very efficient conversion of unstructured speech data to a structured and searchable text representation.

Here is a short clip of Veridict’s live transcription technology, demonstrating incremental live transcription (in Swedish).

For addiotional information, please contact Veridict.