Audio thumbnails for spoken content without transcription based on a maximum motif coverage criterion

International audienceThe paper presents a system to create audio thumbnails of spo- ken content, i.e., short audio summaries representative of the entire content, without resorting to a lexical representation. As an alternative to searching for relevant words and phrases in a transcript, unsupervised motif discovery is used to find short, word-like, repeating fragments at the signal level without acous- tic models. The output of the word discovery algorithm is ex- ploited via a maximum motif coverage criterion to generate a thumbnail in an extractive manner. A limited number of relevant segments are chosen within the data so as to include the maxi- mum number of motifs while remaining short enough and intel- ligible. Evaluation is performed on broadcast news reports with a panel of human listeners judging the quality of the thumb- nails. Results indicate that motif-based thumbnails stand be- tween random thumbnails and ASR-based keywords, however still far behind thumbnails and keywords humanly authored