Experimental frameworks

• Plot WRR(SNR), for:

  • audio-only speech recognition;
  • visual-only speech recognition;
  • audio-visual speech recognition

SNR = S/N = signal-to-noise ratio

• Confusion matrices:

Department of Informatics

Aristotle University of Thessaloniki

• Measures for performance evaluation:

Previous slide

Next slide

Back to first slide

View graphic version