Automating Therapy for Improving Speech of Patients With Apraxia

We need to develop an automated process for providing therapy to improve speech of patients who are suffering from some kind of hearing disability or speech apraxia. The system will be model based speech recognition and employ speech visualization for giving the patient the feedback of his own speech compared to the standardized sample. Using various voice samples various graphs are generated. Once the samples are generated and analysed the patient’s voice can be easily graded comparing it against the ideal sample. Thereby giving us a medium to provide a feedback for the voice sample. A score is generated depicting the similarity between both the samples. An ideal speech has a score of 9. Any score below 5 depicts poor pronunciation or unacceptable results. It is aimed to fully automate the therapy used in the treatment of hearing related disabilities.

Keywords: Speech Recognition, Speech Recognition system, Ideal speech sample, Speech therapy, Speech apraxia, Recorded speech sample, signal graph, frequency graph, coherence

Suggested Citation: Suggested Citation

Bhardwaj, Harshit and Chakarvarti, Mukul and Kumar, Nikhil and Mittal, Shubham, Automating Therapy for Improving Speech of Patients With Apraxia (April 1, 2020). Proceedings of the International Conference on Innovative Computing & Communications (ICICC) 2020, Available at SSRN: https://ssrn.com/abstract=3565915 or http://dx.doi.org/10.2139/ssrn.3565915