Summary of Research on Whispered Speech Recognition Technology

Speech recognition technology has been studied in normal speech and has achieved better recognition performance, but whispered speech recognition still needs a lot of research work. This article briefly introduces the development of whispered speech recognition technology in recent years, summarizes the application and improvement of traditional models in whispered speech recognition, and explains the main technologies, advantages and disadvantages of these methods. Due to the great contributions of deep learning and end-to-end to the field of whispered speech recognition, this article introduces their research results in the field of whispered speech recognition, and explains their main methods, advantages and disadvantages. On this basis, the whispered corpus and pseudo-whispered generation methods are summarized. Finally, the process of whispered speech research was discussed and analyzed objectively.