Work

Tandem decoding of children's speech for keyword detection in a child-robot interaction scenario

Year: 2011

Type: article

Abstract: In this article, we focus on keyword detection in children's speech as it is needed in voice command systems. We use the FAU Aibo Emotion Corpus which contains emotionally colored spontaneous children's speech recorded in a child-robot interaction scenario and investigate various recent keyword spotting techniques. As the principle of bidirectional Long Short-Term Memory (BLSTM) is known to be well-suited for context-sensitive phoneme prediction, we incorporate a BLSTM network into a Tandem model for flexible coarticulation modeling in children's speech. Our experiments reveal that the Tandem model prevails over a triphone-based Hidden Markov Model approach.

Source: ACM Transactions on Speech and Language Processing

DOI: 10.1145/1998384.1998386

Authors: Martin Wöllmer, Björn Schuller, Anton Batliner, Stefan Steidl, Dino Seppi

Institutions: Technical University of Munich, Friedrich-Alexander-Universität Erlangen-Nürnberg, KU Leuven

Language: English

Topic: Speech Recognition and Synthesis

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: green

Funder: Seventh Framework Programme

Award: G4217892124