Structured Course Generation for Kinect-based Chinese Character Learning: A Prerequisite-Driven Approach for Improved Learning

Abstract

Learning Chinese characters is challenging for non-native speakers due to their complex visual structures and weak phonetic correspondence. Although gesturebased interactive systems have been proposed to enhance engagement, they often lack structured course plans, leading to cognitive overload and inefficient learning. To address this limitation, this paper proposes a systematic course generation method for a Kinect-based Chinese character educational system. The method first decomposes characters into multi-granularity learning objects. It then constructs a prerequisite structure based on inclusion relationships and assigns difficulty levels according to structural complexity. An easy-to-complex learning path is subsequently generated using topological sorting. To evaluate the effectiveness of the proposed method, a controlled user study involving 40 non-native learners is conducted, which compares a guided learning mode (Treatment) with an unguided mode (Control). The results demonstrate that the Treatment group significantly outperforms the Control group in written performance, perceived cognitive load, system usefulness, and learning satisfaction (P < 0.05). These findings demonstrate that the proposed method improves both learning effectiveness and user experience, and highlight the value of structured learning path design in gesture-based e-learning systems.