Research on Human-Computer Interactive Children's Emotional Recognition Technology Integrating Attention Mechanism
With the development of artificial intelligence,more and more intelligent devices are being applied to the education of preschool children.Although emotion recognition technology has made significant progress in many aspects,there are still some chal-lenges.One of them is how to better adapt to specific user groups,especially for emotional recognition of children.Children's emo-tional expression may be more complex,so more meticulous technical means are needed for effective emotional recognition.This study constructs a bimodal emotion recognition model using Mel spectrograms and word embedding vectors of text information.The experi-mental results show that when the dataset size is 500,the proposed algorithm model has a recall rate of 0.91 and an F1 value of 0.92.The loss function value is 0.12,with an accuracy of 0.91.The research results indicate that the proposed method has good perform-ance in children's emotional recognition.
human-computer interactionpreschool educationemotion recognitionmel frequency cepstrum coefficient
万方数据
维普
