Text-Independent Speaker Identification Using Deep Learning Model of Convolution Neural Network

Abstract

Speaker recognition approach can be categorized into speaker identification and speaker verification. These two subfields have a bit varied in definition from domain usage. If we has a voice input, the goal of speaker verification is for authentication by determining an answer from a question: "is the voice someone's voice?" For speaker identification, will try to find an answer: "the voice is whose voice?" It can be thought that verification is a special case of open-set identification. In this work, deep learning model using a convolution neural network (CNN) for speaker identification is proposed. The voice input to the method is no constrained on the words the speaker speaks. That means it is in a form of text-independent of which more difficult than text-dependent system. By the method, each 2 seconds of the speaker voice is transform to a spectrogram image and input to the generated CNN model training from scratch. The proposed CNN based method is compared to the classic Mel-frequency cepstral coefficients (MFCCs) based featured extraction method classified by support vector machine (SVM). Where, up to date, MFCC is the most popular feature extracted method for audio and speech signal. Our proposed method that the spectrogram image is used as an input is also compared to a case when image of raw signal wave is employed to the CNN model. Experiments are conducted on the speech from five speakers speak in Thai language of which voices are extracted from YouTube. It reveals the proposed CNN based method trains on spectrogram image of voice is the best compared to the other two methods. The average classification results of the testing set by the proposed method is 95.83%. For MFCC based method is 91.26% and for CNN model trained on image of raw signal wave is only 49.77%. The proposed method is very efficient when only short utterance of voice is used as an input.Index Terms-Convolution neural network (CNN), deep learning, speaker recognition, speaker identification, text-independent.

Hybrid machine learning classification scheme for speaker identification

“…The accuracy of the proposed hybrid scheme for the MFCC feature is 2% superior to the CNN scheme proposed by Bunrit et al [41]. These findings show that our proposed features provide the highest accuracy for all of the classifiers considered in this study.…”

“…The accuracy of the proposed hybrid scheme for the MFCC feature is 2% superior to the CNN scheme proposed by Bunrit et al [41].…”

“…From the above table, observed results show that the proposed RF‐SVM classifier model produces a higher average kappa metric (99%) as well as f1_score (92.4%) and thus offers improved speaker identification. Also, the kappa and f1_score measures of CNN (Bunrit et al [41]) using spectrogram feature are 98.27% and 91.57%, respectively, which is 1% lower than the proposed hybrid approach using combined spectral features.…”

“…This work identifies the unknown test speaker using the trained speaker models along with the classifier RF‐SVM. Various traditional techniques, such as Gaussian mixture model (GMM), SVM, AFB‐based SVNN [22], deep neural network (DNN) [40], and CNN [41], are evaluated against the RF‐SVM classifier to measure the efficiency of the recommended scheme in speaker identification. This paper combines traditional datasets with noise from additional sources to train classifier algorithms.…”

“…Note that RF‐SVM had superior identification rates: An increase of about 13% more than GMM and 2% greater than DNN. From the performance analysis, the proposed combined spectral feature‐based hybrid RF‐SVM scheme produces average accuracy of 99.04% when compared to CNN‐based method using spectrogram image feature got the average accuracy of 95.83% (Bunrit et al [41]) and i‐vector scheme (Dehak et al [19]). The results in Figure 6 indicate that the proposed scheme is more appropriate for recognizing speakers.…”