@article{YANG2026, 
author = {Yi YANG and Feng QIU},
title = {Design, Implementation, and Functions of the Multimodal Analytical Framework for Teacher Discourse},
year = {2026},
journal = {Modern Educational Technology},
volume = {36},
number = {2},
pages = {51-59},
keywords = {teacher discourse, multimodal analysis, expression recognition, posture recognition, discourse intention},
url = {https://www.sciopen.com/article/10.3969/j.issn.1009-8097.2026.02.006},
doi = {10.3969/j.issn.1009-8097.2026.02.006},
abstract = {Teacher discourse is the mixed use of clues such as text, rhythm, expression, and action, and is characterized by multimodality. To comprehensively analyze teacher multimodal discourse and facilitate effective interactive teaching, this paper firstly designed the multimodal analytical framework for teacher discourse centered on intention expression based on Halliday’s systemic functional linguistics theory and Gu Yueguo’s conceptual model. Subsequently, the technical implementation of this framework was discussed, employing the mechanism for teacher’s expression recognition based on CNN model and the mechanism for teacher’s posture recognition based on PoseNet model, to achieve objective and accurate identification of teacher’s expressions and postures. Further, a “four-digit encoding” rule for teacher’s actions was designed. Finally, the main functions of the multimodal analytical framework for teacher discourse were expounded, which included reflecting teacher’s emotional states and instructional style variations, presenting the underlying discourse intentions of the corpus from a micro perspective in a three-dimensional manner, and using multimodal data to assist new teachers in improving their teaching skills. The analytical framework designed in this paper provided important ideas for studying teacher discourse intentions, while its implementation mechanism and functional design offered operational guidelines for supporting teacher growth.}
}