A Hybrid Approach to Gender Classification using Speech Signal

Pir, M. Yasin; Wani, Mohamad Idris

doi:10.32628/ijsrset196110

Cited by 4 publications

(2 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The classification of gender and age from direct speech signals can be logically related to the time or frequency domains. In the time domain analysis, we directly measure the speech signals considering the content of a signal for evaluating information regarding a speaker; on the other hand, in the frequency domain analysis, the frequency content of a speech signal is used to form a spectrum for evaluating information regarding a speaker, which is analyzed accordingly [ 17 , 18 ]. The smart gender-age recognition system used the main variation in the levels of power and the frequency content of the two genders to identify the gender-based speech signals [ 9 , 10 ].…”

Section: Literature Reviewmentioning

confidence: 99%

Age and Gender Recognition Using a Convolutional Neural Network with a Specially Designed Multi-Attention Module through Speech Spectrograms

Tursunov

Mustaqeem

Choeh

et al. 2021

Sensors

View full text Add to dashboard Cite

Speech signals are being used as a primary input source in human–computer interaction (HCI) to develop several applications, such as automatic speech recognition (ASR), speech emotion recognition (SER), gender, and age recognition. Classifying speakers according to their age and gender is a challenging task in speech processing owing to the disability of the current methods of extracting salient high-level speech features and classification models. To address these problems, we introduce a novel end-to-end age and gender recognition convolutional neural network (CNN) with a specially designed multi-attention module (MAM) from speech signals. Our proposed model uses MAM to extract spatial and temporal salient features from the input data effectively. The MAM mechanism uses a rectangular shape filter as a kernel in convolution layers and comprises two separate time and frequency attention mechanisms. The time attention branch learns to detect temporal cues, whereas the frequency attention module extracts the most relevant features to the target by focusing on the spatial frequency features. The combination of the two extracted spatial and temporal features complements one another and provide high performance in terms of age and gender classification. The proposed age and gender classification system was tested using the Common Voice and locally developed Korean speech recognition datasets. Our suggested model achieved 96%, 73%, and 76% accuracy scores for gender, age, and age-gender classification, respectively, using the Common Voice dataset. The Korean speech recognition dataset results were 97%, 97%, and 90% for gender, age, and age-gender recognition, respectively. The prediction performance of our proposed model, which was obtained in the experiments, demonstrated the superiority and robustness of the tasks regarding age, gender, and age-gender recognition from speech signals.

show abstract

Section: Literature Reviewmentioning

confidence: 99%

Age and Gender Recognition Using a Convolutional Neural Network with a Specially Designed Multi-Attention Module through Speech Spectrograms

Tursunov

Mustaqeem

Choeh

et al. 2021

Sensors

View full text Add to dashboard Cite

show abstract

“…Manjunath et al [19] presented an acoustic event classification model using new features extracted from spectrogram blocks. A hybrid model using 1-D Stationary Wavelet Transform (SWT) for signal denoising and reconstruction was given to ANN for gender recognition developed by Yasin et al [20]. Anna V. Kuchebo et al [21] presented a deep learning network to classify gender from speech using the Mel spectrogram of speech of the Mozilla voice dataset.…”

Section: Introductionmentioning

confidence: 99%

Voice-Based Gender Recognition Model Using FRT and Light GBM

2023

Teh. vjesn.

View full text Add to dashboard Cite

Voice-based gender recognition is vital in many computer-aided voice analysis applications like Human-Computer Interaction, fraudulent call identification, etc. A powerful feature is needed for training the machine learning model to discriminate a gender as male or female from a voice signal. This work proposes the use of a gradient boosting model in conjunction with a novel Cumulative Point Index (CPI) feature computed by Forward Rajan Transform (FRT) for gender recognition from voice signals. Firstly, voice signals are preprocessed to remove the nonsignificant silence period and are further framed and windowed to make them stationary. Then CPI is computed using the first coefficients of FRT and concatenated to form a feature set, and it is used to train the Light Gradient Boosting Machine (LightGBM) to recognize the gender. This approach provides better accuracy and faster training compared with the state of the art techniques. Experimental results show the primacy of the FRTCPI over other standard features used in the literature. It is also shown that the proposed features, in combination with LightGBM, provide better accuracy of 95.26% with a less computational time of 2.25 s over the challenging large datasets like Speech Accent Archive,

show abstract