Speech perception based on mapping speech to image by using convolution neural network

摘要

Although speech perception has been studied for more than sixty years and a great deal about how the system works has been researched, there is still more to be discovered. Previous research considered speech perception in only one aspect of sound such as the speech recognition problem. Speech perception focuses on the process that operates to decode speech sounds no matter what words those sounds might comprise. In this paper, we introduce a new approach for the speech perception as learning the map between speech and other information that we receive from the surrounding environment vice other sensation. So the speech perception problem would be turn into the relationship learning. In this model, we propose using a convolution neural network to map speech signal to image.