About this document
Stutter-TTS: Enhancing Stuttered Speech Recognition by vijayalakshmis is a document available to read on EtoBox.
The document introduces Stutter-TTS, a neural text-to-speech model designed to synthesize various types of stuttered speech, addressing the challenges faced by automatic speech recognition (ASR) systems in understanding stuttered utterances. By incorporating special tokens into the training data to represent stuttering characteristics, the model achieves high accuracy in synthesizing stutter events and demonstrates a 5.7% relative reduction in word error for stuttered utterances when fine-tuning an ASR mode
- Author
- vijayalakshmis
- Language
- EN