Lip Feature Extraction and Classification Algorithm in English Speech Recognition

Abstract

In view of the problems existing in the existing lip feature extraction algorithms, such as large calculation, too much manual intervention in the process of processing, and poor practicability, the existing lip feature information extraction schemes are compared and analyzed. The lip region image is extracted by feature extraction including geometric features and Gabor features. These features are evenly sampled between feature extraction, and Gabor feature extraction is performed. At the same time, support vector machine is also studied and analyzed. Using radial basis function and one-to-one voting classifier to classify the geometric features and Gabor features extracted previously, and combining the recognized face features, an adaptive weight allocation rule based on recognition rate is proposed. Finally, the proposed classification of facial expression features in the lip region is fused with the overall facial expression features, and compared with the traditional facial expression recognition algorithm, which proves the importance and effectiveness of the algorithm.