Deep Convolutional Neural Networks for Intelligent Emotional Computing: A Multidimensional Analysis and Practical Application of Emotional Features in Digital Learning Interfaces

This study explores an intelligent emotional computing method baed on deep convolutional neural networks (DCNNs) aimed at accurately identifying and analyzing learners' emotional states within digital learning interfaces. A multi-channel, multi-scale convolutional neural network structure is designed to effectively integrate multimodal emotional features such as facial expressions, physiological signals, behavioral data, and learning environments. Experimental validation on both self-built datasets and publicly available datasets demonstrates that the proposed model achieves superior performance in emotion recognition tasks, with an average accuracy rate of 85.2^%, significantly outperforming traditional methods based on handcrafted features and shallow networks.