CNN-Driven Innovations in Dysarthria Classification and Speech Disorder Management

This research article addresses dysarthria, a speech disorder frequently observed in individuals with cerebral palsy (CP) or amyotrophic lateral sclerosis (ALS). Early detection of dysarthria is essential for effective treatment and improved patient outcomes, as it significantly impacts communication abilities. Utilizing the TORGO database, which comprises 2000 speech samples from both dysarthric and non-dysarthric individuals across various genders, this study employs a Convolutional Neural Network (CNN) for classification. The dataset includes 500 samples each from both groups, collected in different sessions. The CNN approach achieves an impressive 96% accuracy in distinguishing dysarthric from non-dysarthric speech, showcasing the potential of deep learning technology in facilitating early dysarthria diagnosis and enabling timely interventions. This study makes substantial contributions to enhancing diagnostic capabilities and offers a promising avenue for improving the quality of life for individuals with speech disorders.