Automatic turn segmentation for Movie & TV subtitles
- Resource Type
- Conference
- Authors
- Lison, Pierre; Meena, Raveesh
- Source
- 2016 IEEE Spoken Language Technology Workshop (SLT) Spoken Language Technology Workshop (SLT), 2016 IEEE. :245-252 Dec, 2016
- Subject
- Signal Processing and Analysis
Motion pictures
TV
Pragmatics
Silicon
Visualization
Timing
Feature extraction
- Language
Movie and TV subtitles contain large amounts of conversational material, but lack an explicit turn structure. This paper present a data-driven approach to the segmentation of subtitles into dialogue turns. Training data is first extracted by aligning subtitles with transcripts in order to obtain speaker labels. This data is then used to build a classifier whose task is to determine whether two consecutive sentences are part of the same dialogue turn. The approach relies on linguistic, visual and timing features extracted from the subtitles themselves and does not require access to the audiovisual material - although speaker diarization can be exploited when audio data is available. The approach also exploits alignments with related subtitles in other languages to further improve the classification performance. The classifier achieves an accuracy of 78 % on a held-out test set. A follow-up annotation experiment demonstrates that this task is also difficult for human annotators.