LARYNX DISEASE ANALYSIS USING DYNAMIC TIME WARPING ALGORITHM?

There are several diseases which cause changes in speech. Larynx cancer is one among them. Cancer can be best diagnosed in a non-invasive method and treated in the early stages. Acoustic voice analysis is an effective and non-invasive method for evaluation and detection of laryngeal pathologies. In this paper we have collected a wide variety of voice samples, including sustained vowels, words, and sentences compiled from a set of speaking exercises for people with laryngeal diseases. We are using multiple sound recordings of a single subject and find out the central tendency and dispersion metrics which improves generalization of the predictive model. Dynamic Time Wrapping algorithm is used to find out the similarity between two voice samples.