Fig 1.
Training and testing stages of an ASI system.
Fig 2.
Features from different transforms.
Fig 3.
Wavelet transform.
Fig 4.
MFCCs extraction.
Fig 5.
Computation of the net activation.
Fig 6.
Conventional speaker identification system.
Fig 7.
Proposed speaker identification in the presence of reverberation.
Fig 8.
Magnitude and phase responses of a comb filter.
Fig 9.
Comb filter and its output.
Fig 10.
Proposed cancelable speaker system.
Fig 11.
Cancelable speaker identification in the presence of reverberation.
Table 1.
Speech database description and ANN parameters.
Fig 12.
Variation of the output recognition rate of the speaker identification system with SNR for different feature extraction techniques without reverberation effect.
Fig 13.
Recognition rate variation in the presence of reverberation, as influenced by diverse feature extraction techniques across different SNR levels.
Fig 14.
Variation of the output recognition rate of cancelable speaker identification system with SNR for different feature extraction techniques without reverberation effect.
Fig 15.
Recognition rate variation in the cancelable speaker identification system under reverberation, across different feature extraction techniques and SNR levels.
Table 2.
Number of epochs required for training the ANN.
Table 3.
Output recognition rates of the speaker identification system for different feature extraction techniques at different SNRs without reverberation effect.
Table 4.
Output recognition rates of the speaker identification system for different feature extraction techniques at different SNRs in the presence of reverberation.
Table 5.
Output recognition rates of cancelable speaker identification system for different feature extraction techniques at different SNRs.
Table 6.
Output recognition rates of the cancelable speaker identification system in the presence of reverberation for different feature extraction techniques at different SNRs.