CNN AND LSTM FOR THE CLASSIFICATION OF PARKINSON'S DISEASE BASED ON THE GTCC AND MFCC

Parkinson's disease is a recognizable clinical syndrome with variety of causes and presentations; it represents rapidly growing neurodegenerative disorder. Since about 90 percent sufferers have some form early speech impairment, recent studies on tele diagnosis focused the recognition voice impairments from vowel phonations or subjects' discourse. In this paper, we present new approach for detection sounds that are based CNN LSTM uses two categories characteristics Mel Frequency Cepstral Coefficients (MFCC) Gammatone (GTCC) obtained noise-removed signals comparative EMD-DWT DWT-EMD analysis. The proposed model divided into three stages. first step, noise removed using methods. second GTCC MFCC extracted enhanced audio signals. classification process carried out in third step by feeding these features models, which designed to define sequential information features. experiments performed PC-GITA Sakar datasets 10-fold cross validation method, highest accuracy dataset reached 100% both EMD-DWT-GTCC-CNN DWT-EMD-GTCC-CNN, dataset, 96.55% DWT-EMD-GTCC-CNN. results study indicate more appropriate accurate assessment PD than MFCC.