A Review: Person Recognition Based on Humming
Nikunj Patel, Vinesh Kapadia, Kinnal Dhameliya, Ninad Bhatt · 2015
The aim of this paper is to provide an overview of different feature extraction techniques for humming based person recognition. Normally, speech is used as an input to the biometric system for recognition of a speaker, but over here, instead of using speech, hum of a person is used for the same purpose. Hum is a sound produced by the nose, in which the oral cavity is closed and the vocal tract is coupled with the nasal cavity. Humming is also more universally available on everyone than speech so it is also applicable for a deaf person as well as an infant with disorder in speech production mechanism. Different features like Mel Frequency Cepstral Coefficients (MFCC), Linear Predictive coefficients (LPC), Perceptual Linear Prediction (PLP) and all other feature vectors are used as an input to the classifier of Gaussian Mixture Model-Universal background Model (GMM-UBM) and polynomial classifier of 2 and 3 order. Measurement of a performance of a system is done based on the Identification/Success Rate and Equal Error rate (EER). Keywords—Speaker recognition, Hum, LPC, MFCC, polynomial classifier, GMM-UBM, EER.