Research on Multimodal Knowledge Points Segmentation Tool for Classroom Teaching Videos

Knowledge points represent the fundamental units of teaching information or concepts and serve as the foundation for teaching content. The segmentation of knowledge points from extensive classroom teaching videos by integrating artificial intelligence technology can assist educators comprehend the way in which course content is presented and teaching methods are employed. However, due to the complexity of teaching scenario and the fact that knowledge points presentations contain multimodal information, there have been few studies conducted to analyze the correlation between the knowledge points and the multimodal features and to segment the video clips from the perspective of the knowledge points. In light of the above, this paper proposes a novel multimodal knowledge points segmentation tool for classroom teaching videos. The tool has three main functions, which are designed to achieve intelligent segmentation and extraction of knowledge points. These include extraction of candidate knowledge points sequence fusing auditory and textual features, extraction of candidate knowledge points sequence based on visual features and segmentation of knowledge points integrating multimodal information. It provides teachers with a convenient platform and method for segmenting knowledge points, thus enabling them to conduct a series of teaching analyses.