Stuttering Sub Challenge

The Stuttering Sub-Challenge focuses on developing automated methods for detecting and recognizing stuttered speech, aiming to improve diagnosis and treatment of this prevalent speech disorder. Current research emphasizes the use of deep learning models, particularly those leveraging self-supervised learning and multi-modal data (combining audio and visual information), to enhance accuracy and robustness in detecting various types of dysfluencies. These advancements hold significant promise for streamlining clinical assessment, enabling earlier intervention, and improving the accessibility of speech recognition technologies for individuals who stutter.