Study Of Mouth Recognition Based On Deep Learning Method

In speech recognition,mouth recognition technology can play a good supporting role.The technology can combine the information of visual channel and auditory channel,and has better recogntion effect compared with the use of separate auditory channels.In this paper,the main problems in the mouth recognition system,such as face detection,human eye detection,lip segmentation and mouth recognition,are studied.This paper presents a method based on the composite features of lip(geometric and pixel feature)to get the accurate positioning of lip gradually,the first step is to use the specific geometric position of facial features,then use pixel feature of lip in HSV color space to get the lip accurately,and this method can be applied to different illumination conditions.Finally,deep learning method is used to extract the features of the lip,and use the self built Chinese mouth database to train the model to complete the task of mouth recognition.The experimental results show that the accuracy of this method is very high and can reach 94%.