Research And Application Development Of Human Lip Contour Recognition Algorithm Based On Deep Learning

At present,deep learning technology has been applied to all walks of life,especially in the fields of image recognition,biological identity authentication and intelligent speech.In the field of biological identity recognition and intelligent speech,there are requirements for lip recognition and detection.At present,most of the lip detection uses lip color space characteristics and lip shape model methods.These methods are easily affected by light and lip deformation,and the recognition accuracy is not stable.In this paper,a method of lip detection and recognition using deep learning technology is proposed,and analyzes its recognition results and accuracy,and implements a mobile app that uses deep learning models for lip detection and beauty makeup.Through the application of the algorithm model proposed in this paper,the recognition and detection results with higher accuracy are obtained.The research results of this paper provide valuable reference for how to use deep learning for lip detection and recognition.This paper introduces the background and research significance of lip detection,and attempts to carry out lip detection based on deep learning method.On the basis of studying the BiSeNet network model,a Res-BiSeNet network model for lip recognition was proposed.The experimental data show that the MIoU of the model on the test set reaches 89%,which can meet the needs of lip detection speed and accuracy in most scenarios and environments.Based on this algorithm,this paper designs and implements an APP for lip beauty makeup on iOS platform,which achieves a high precision detection effect.The research work in this paper has verified the feasibility and effectiveness of deep learning technology in lip detection tasks.The lip beauty APP developed on iOS platform provides a good reference for this kind of research and application.