Research on the Strategies of Emotional Interaction between Teachers and Students in Online Courses

You can refer to this information. It should be modifiable!

Emotional interaction: When a computer plays a multimedia program, the programmer can give instructions to control the operation of the program with his own thoughts and emotions, instead of unilaterally executing the program. The program responds accordingly after receiving the corresponding instructions from the programmer. This process and behavior is called emotional interaction.

Emotional computing and artificial psychology are new research directions in the fields of harmonious human-computer interaction and artificial intelligence, and they are also new intersections of mathematics, information science, intelligent science, neuroscience, physiology and psychological science. The research of affective computing is trying to create a computing system that can perceive, recognize and understand people's emotions and make intelligent, sensitive and friendly responses to people's emotions. Under the guidance of artificial psychology and affective computing theory, this paper makes an exploratory study on the techniques and methods of human-computer emotional interaction, and discusses the related technologies of human-computer emotional interaction, such as emotional modeling, expression recognition and expression synthesis, so as to provide relevant technical support for building a harmonious man-machine. The main research contents and innovations of this paper are as follows: (1) A facial expression recognition method based on the combination of Zernike moments and HMM is proposed. In the aspect of feature extraction, the moment feature extraction method based on local face information is adopted. By segmenting the local face region, the eye and mouth sub-images which can best reflect the expression information are extracted, and the moment feature vectors of the sub-images are calculated. On the classification problem, HMM based on MCE training criterion is used to classify. Experiments show that the definition of the criterion function is more reasonable than the original criterion function, and it can effectively use the discriminant information in the training sample set, make full use of the training data, and improve the performance of HMM. (2) Emotional computing model is considered as a key component to achieve more effective human-computer interaction through emotional interaction with users. In the process of establishing emotional computing model, a numerical classification and description method of emotion is proposed, and the numerical space of emotion is established by combining dimension theory and basic emotion theory. The numerical description method of emotional components such as personality, emotion and demand and the dynamic processing method of demand are put forward, and the mapping relationship between emotional components is established. How to effectively organize and coordinate various emotional components and realize the mapping from external environment and emotional state to emotional behavior is a difficult problem in emotional modeling. In order to solve this problem, an emotional computing model based on emotional cognitive evaluation theory is proposed. This model uses Markov process to describe the change and transfer process of emotional state, and uses markov decision processes to establish the relationship between emotional components and emotional expression. (3) In the part of facial expression synthesis, Candide model is used as a parameterized facial model. After texture mapping of the mesh, the facial expression of the facial model is controlled by controlling AU(Action Units), so as to achieve the purpose of parameterized control, and then a variety of expressions can be accurately synthesized. An automatic interactive model adjustment algorithm is proposed. When the automatic extraction effect is not good, we can interactively modify the front features and estimate the depth of feature points, and adjust the non-feature points of the model by radial basis function interpolation. Compared with other methods, the algorithm realizes adaptive adjustment under the condition of good image quality, and has certain practical value. Based on the above algorithm and model, an anthropomorphic network information consulting service system is designed and implemented, and the validity and correctness of the model are verified. The emotional model in this paper enables the information consulting service system to show rich emotions. Compared with the general information consulting service system, the information consulting service system is more intelligent and anthropomorphic.