Lip-reading via a DNN-HMM Hybrid System Using Combination of The Image-based and Model-based Features

Introducing features that better represent the visual information of speakers during the speech production is still an open issue that highly affects the quality of the lip-reading and Audio Visual Speech Recognition (AVSR) tasks. In this paper, three different types of visual features from both the image-based and model-based ones are investigated inside a professional lip reading task. The simple raw gray level information of the lips Region of Interest (ROI), the geometric representation of lips shape and the Deep Bottle-neck Features (DBNFs) extracted from a 6-layer Deep Auto-encoder Neural Network (DANN) are three valuable feature sets compared while employed for the lip reading purpose. Two different recognition systems, including the conventional GMM-HMM and the state-of-the-art DNN-HMM hybrid, are utilized to perform an isolated and connected digit recognition task. The results indicate that the high level information extracted from deep layers of the lips ROI can represent the visual modality with advantage of "high amount of information in a low dimension feature vector". Moreover, the DBNFs showed a relative improvement with an average of 15.4% in comparison to the shape features and the shape features showed a relative improvement with an average of 20.4% in comparison to the ROI features over the test data.