A Multidimensional Image Feature-Based Method for Classifying Student Group Learning Behaviors and Its Educational Intervention Strategies

With the deep integration of educational informatization and intelligent technology, multidimensional image data collected by smart devices has become a rich resource for analyzing student group learning behaviors.Accurate classification of these behaviors is essential for optimizing teaching strategies and enhancing educational quality.However, existing research faces three major limitations: (1) reliance on single image features, overlooking the association between local structural features such as body movements and complex learning environments; (2) simplistic feature fusion methods that fail to account for the correlation and varying importance of multidimensional features, thereby limiting classification accuracy; and (3) a lack of systematic development of educational intervention strategies, hindering the practical application of behavioral analysis.To address these issues, this study proposes a multidimensional image feature extraction method for classifying student group learning behaviors.The method integrates local structural features, globally weighted local phase quantization (LPQ) index structure features, scene features, color features, and image information entropy to construct a comprehensive feature representation framework.A high-efficiency classification model is developed in tandem with targeted educational intervention strategies, forming a complete framework of "feature extractionbehavior classification-intervention implementation."The research outcomes are expected to significantly improve the accuracy of learning behavior classification, provide robust data support for personalized teaching and optimized learning guidance, and promote the deep integration of behavior analysis techniques with practical educational interventions in the context of educational informatization.