Research On Facial Micro-expression Recognition For Online Learning

With the continuous empowerment of Internet technology,the education industry is deeply integrated with the Internet.Online learning is an open learning environment based on Internet technology and education platforms.Unlike traditional classrooms,online learning provides learners with a more convenient way to learn,in contras,there is a serious problem in online learning:teachers can't sense students'learning emotions in real time.In order to improve the learning efficiency of the learner,the educator must grasp the current learning state and learning mood ofthe learner as soon as possible to adjust in time.However,learners may try to hide emotions during the learning process and subtle expression changes are instantaneous,which traditional fecial recognition techniques are difficult to capture.Therefore,it is especially important to apply facial micro-expression recognition technology to online learning.In the micro-expression recognition process,the extracted feature quality plays a decisive role in the accuracy of recognition.Among the commonly used feature extraction methods are optical flow method,3D histogram,LBP_TOP,etc.,among them,LBP-TOP algorithm is a better algorithm based on video micro-expression recognition.However,the LBP algorithm which compares the size of the central pixel and the neighboring pixel is not sensitive to illumination as well as is greatly affected by noise.Therefore,this paper improves a micro-expression recognition method based on optical flow method and LBP-TOP feature fusion.The specific work of this paper is as follows:(1)Improve and compare micro-expression recognition algorithm:two classical microfacial feature extraction algorithms and microfacial classification technology are introduced.Because of the limitation of single feature extraction algorithm,this paper proposes a method of micro-expression recognition based on optical flow method and LBP-TOP feature fusion,then compares the two algorithms based on the same reference database and the same expression classification technology;(2)Creation of micro-expression database:based on the online learning mode,the advantages and disadvantages ofthe existing database are analyzed and summarized.The micro-expression video of the tester is obtained based on the creation principle of the CASMEII database and the obtained video is processed by image processing including video frame extraction,image preprocessing,face detection and face segmentation;Finally,micro-expressions are applied to the processed images to complete an application scenario database suitable for online learning;(3)Application of online learning field:introduce the definition and structure of the online learning system and analyze the shortcomings of the existing online learning system structure.Based on the online learning platform,the emotional state recognition module is designed and implemented.The micro-expression recognition technology and micro-feature feature matching are used to feedback the learner's emotional state in real time during the teaching process.According to the feedback result,the teaching process is dynamically adjusted and improved.The phenomenon of "lack of emotion" in online learning helps to improve the learning efficiency of learners.