Traditional systems of speech recognition based interaction need to identify the accurate semantics and grammar in the speech to be recognized. This kind of systems always contain problems such as using a specific language, presetting voice template, they are not suitable for speech recognition interaction in augmented reality children’s books. For these reasons, this paper studies the pronunciation characteristics and interactive needs of children, analyses the differences among template matching method, random model method and artificial neural network method. Then uses template matching method to design a speech recognition system for small vocabulary that can replace the voice template, and applies it to the speech recognition interaction in an augmented reality children's reading application. Finally, the effectiveness of the proposed method is verified through experiments, which provides a new direction for the application of speech recognition technology in augmented reality children’s books.
[1]
Shrikanth S. Narayanan,et al.
Analysis of children's speech: duration, pitch and formants
,
1997,
EUROSPEECH.
[2]
Alberto Del Bimbo,et al.
A Natural and Immersive Virtual Interface for the Surgical Safety Checklist Training
,
2014,
SeriousGames '14.
[3]
Yang Li,et al.
Gestures without libraries, toolkits or training: a $1 recognizer for user interface prototypes
,
2007,
UIST.
[4]
Radu-Daniel Vatavu,et al.
Gestures as point clouds: a $P recognizer for user interface prototypes
,
2012,
ICMI '12.