A voice activity detector using SVM and Naïve Bayes classification algorithm

Abstract

Voice or Speech Pathology analysis performs a significant role in the recent record of medical experts. The need for research is the recognition and classifications of tone of pathological voices are believed as a challenging work in the field of speech analysis still now. Commonly Patients are in a position to identify a change in voice parameters, such as hoarseness; however the voice pathologies can result from a wide spectral range of causes, like common cold to a malicious tumor. Medical experts like otolaryngologists were discovering a genuine quantity and range of speech pathologies from the Patients conversation. Unluckily, the current classification rate of voice pathology by the human experts is merely about 60–70%. Thus tone of voice or speech pathologies can be found by the endoscopy techniques and strategies like laryngostroboscopy or medical micro laryngoscopy, which distress the individual to a great scope which is expensive also. The primary objective of the research work is to assist this speech pathology finding process with computer structured diagnostic tools. This speech pathology diagnosis system works predicated on the support of the medical clinic based mostly professional otolaryngologists, by determining and figuring out the chance of the pathology automatically without the endoscopy which escalates the detection of speech pathology at the initial stage. In this research work, the conversation signal is examined by the acoustic guidelines and variables like transmission energy, pitch, Silence removal, Windowing, Mel consistency and occurrence Cepstrum, and Jitter. At the final end, the classification strategy i.e Support Vector Machine is employed to classify the standard and pathology speech, predicated on the features extracted in the last phase. Predicated on the results & conversation and dialogue pointed out below, thus the Speech pathology recognition system successfully categorized and labelled the normal tone of voice and the pathological speech which assists with diagnosing and examining the patient.