Gender classification using pitch and formants

Abstract

A gender classification system is proposed based on pitch, formants and combination of both. Ten Hindi digits database has been prepared for fifty speakers. Each Speaker has spoken each digit ten times. Formants derived from speech samples have been used for gender classification. Gender classification has been also done by using pitch extracted from different methods. Autocorrelation, Cepstrum and Average Magnitude Difference (AMDF) methods have been used for pitch determination from speech samples. Formants in combination with pitch are also used for gender classification. A feature vector consisting of pitches derived from all the above mentioned pitch determination methods was also used for gender classification. Experiments were performed for both open-set and closed-set gender classification. Autocorrelation method performed best for gender classification in open-set. Hybrid method (Autocorrelation +AMDF+ Cepstrum) performed best for gender classification in closed-set.

Mel-frequencies Stochastic Model for Gender Classification based on Pitch and Formant

“…AMDF [2] 74,44% Vector quantization [2] 83.5% Euclidean distance [3] 97.05% MFCC [4] 98.8% FFT [5] 80% Neurofuzzy [6] 57.5% Stochastic model 94.11%…”

“…A vector quantization method also used as the method of gender identification. The average of accuracy using these method is 83.5% [2]. A Nearest neighbor is other method to identify the gender of speaker.…”

Categorical perception of pitch: Influence of language tone, linguistic meaning, and pitch contour

“…[1] The quality of the sound is dependent on the highness or lowness of tone pitch along with other acoustic cues. It is a major cue for the identification of suprasegmental aspects of speech; [2] helpful in gender identification; [3,4] age identification; [5,6] emotional arousal; [7] cultural variations; [8] and sociocultural aspects. [9] The change in pitch does not vary the meaning of spoken words or sentences in most of the Indo-European languages such as English, Hindi, and Sanskrit, but in Sino-Tibetan languages such as Mandarin, Cantonese, and Thai, the change in the pitch contours changes the meaning of the words.…”

Gender Classification from Facial Images

“…Only those algorithms that predicted the genders from facial images are selected, and not those based on body [36,37] or gait [38]. Furthermore, gender prediction from speech [39] has also not been included for comparison. The list of datasets used varies from one work to the other.…”