Contrastive Analysis Of Supervised And Unsupervised Learning Techniques For Voice Pathology Detection And Classification

Abstract

The development of technology makes it possible to offer better solutions to the complicated issues that people encounter. The early identification, treatment, and ongoing monitoring provided by today's smart healthcare sectors are crucial in lowering hospital visits, travel expenses, and waiting times.A medical condition known as voice pathology affects the vocal chords and makes it difficult for the patient to speak. As a result of this, the patient may experience difficulty communicating. A study that was only recently presented found that vocal pathology detection systems are capable of accurately diagnosing voice pathologies at an early stage.These systems made use of machine learning strategies, which are regarded as particularly reliable instruments for identifying speech disorders. However, the majority of suggested algorithms for detecting voice disorders used small databases.The low accuracy rate continues to be one of the most difficult problems for these approaches. A technique for identifying voice pathology is described in this research paper.Utilizing the Mel-Frequency Cepstral Coefficient, the voice features are retrieved (MFCC). Vowel /a/ speech samples were equally obtained from the Saarbrücken voice database (SVD). As assessment indices, accuracy is used to compare the effectiveness of various machine learning classifiers. The voice signals in this work are classified as either healthy or disordered using a CNN architecture.