Isolated Words Recognition Algorithm Based on Deep Learning in Heliumspeech Unscrambling

Heliumspeech, language spoken by divers in the deep-sea who breathe high-pressure helium oxygen mixture, is almost unintelligible. This paper focuses on the speech recognition technology of isolated words in heliumspeech. The convolution neural network (CNN), connectionist temporal classification (CTC) and transformer are combined into a speech recognition network to unscramble heliumspeech, and a heliumspeech database is set up to train and test the neural network. Experimental results show that the accuracy of the speech recognition network proposed is up to 91.38%, and the accuracy of each isolated word is more than 90%.