Voice Activity Detection in Noisy Environments Based on Double-Combined Fourier Transform and Line Fitting

Figure 2

State transition model diagram for endpoint detection [17].