Signal Processing and Speech Communication Laboratory
homestudent projects › Wake-up-word recognition

Wake-up-word recognition

Status
Finished
Type
Student Project
Seminar Type
- None -
Announcement date
09 Mar 2015
Student
Clemens Vayda & Wolfgang Hrauda
Mentors
Research Areas

Short Description

In a voice controlled home automation system that is alway listening, an important issue is how to wake-up the system that it listens to a possible command. This wake-up word should be reliably recognize and at the same time the false alarm rate should be as low as possible. In the ideal case the system will also distiguish between the wake-up word uttered with the intention of waking up the system or whether it just appeared in a conversation.

Your Tasks

  • Literature survey for keyword-spotting and wake-up-word recognition
  • Implementation and improvement of promising algorithms
  • Evaluation of the algorithm

Your Requirements

  • Digital Signal Processing
  • Speech Communication 1+2
  • Good MATLAB skills
  • HTK or Kaldi would be nice, but is not necessary

References

[1] V.Z. Këpuska and T.B. Klein: A novel Wake-Up-Word speech recognition system, Wake-Up-Word recognition task, technology and evaluation, in Nonlinear Analysis: Theory, Methods & Applications 71(12):2772–2789, 2009