Design and Preliminary Application of a CV-Based Multimodal Teaching Support System for Higher Vocational Education

Abstract

In recent years, the steady progress of artificial intelligence has brought computer vision (CV) into the spotlight of educational research. Compared with traditional approaches that mainly depend on text or speech, CV is capable of processing multiple information streams—such as images, recognized text, and structural features—and reorganizing them into meaningful teaching resources. This ability makes it possible to support classroom instruction in a way that is more visual, interactive, and efficient. For higher vocational education, where learning tasks are strongly practice-oriented and course materials often contain diagrams or graphical elements, such technologies are particularly relevant. This paper examines how CV-based multimodal techniques can be applied in vocational teaching, with specific attention to resource management, knowledge extraction, and interactive support in the classroom. A prototype framework was designed, integrating functions such as network topology identification, keyword extraction from slides, and the digital capture of handwriting from blackboard work. When tested in teaching scenarios, the system showed promise in improving students’ understanding and participation, while also reducing repetitive tasks for instructors. The study therefore provides both a conceptual foundation and preliminary practical evidence for advancing the digital transformation of vocational education.