Detection and Classification of Kidney Disorders using Deep Learning Method

Abstract

Speech or Voice Pathology investigation performs a significant role in the recent record of Health Industry. The need for analysis is that the detection and classifications of tones of pathological voices, till now, which is believed as a tough task within the sector of speech analysis. Sometimes Patients are probably in tough state to identify a modification in voice parameters, like hoarseness; however the voice pathologies might result from a large spectral fluctuate of causes, like respiratory disease t o a cruel tumor. Medical practitioners like otolaryngologists were discovering different kinds of speech pathologies from the patient‘s speech from mouth i.e. oral communication. Unluckily, this classification rate of Speec h pathology by the physician consultants is just concerning 60-70%. Thus tone of voice or speech pathologies is found by the analysis techniques like laryngostroboscopy or small laryngoscopy, which distress the individual to a decent scope, in addition to that it is expensive. It is not possible to detect the speech pathology at the initial stage by manual diagnosis. The primary objective of this paper is to propose automatic diagnostic tools to assist the voice or speech diagnosis. This speech pathology identification system works based on the support of the medical practitioner, which helps in identifying the pathology even in the beginning stage. In this paper, the speech or voice signal is examined by the acoustic variables like Noise removal, Windowing, Smoothing, Mel consistency and Jitter. Finally the classification of voice pathology is done based on the Neural Networks and Deep Learning. The experimental results were also discussed in detailed manner. From the experimental results it is clear that the Speech pathology recognition s ystem successfully classified and labeled the normal voice and the pathological voice.