Emotion recognition-based framework for supporting English learning environments using BBMO-CNN

Students achieve better results in English language learning when they have strong motivation, self-belief, and a desire to participate while feeling emotionally supported. Emotional health plays a significant role in how students learn and succeed in educational environments. Despite the increasing application of technology in education, emotion detection systems remain insufficiently integrated into flexible and supportive learning environments. This study explores how embedded emotion recognition technology supports English teaching while improving emotionally supportive learning environments. This study uses facial expression metrics to track student emotions, while initial data smoothing is performed using a median filter. For feature extraction, Local Binary Patterns (LBP) is utilized to capture subtle variations in facial expressions. This study introduces a Bumble Bee Mating Optimized Convolutional Neural Network (BBMO-CNN) for emotion recognition through optimized network parameter tuning. The BBMO algorithm is employed to fine-tune the parameters of the CNN, enhancing its ability to accurately classify emotional states from facial expressions. This method improves the overall performance of the emotion recognition system, enabling higher accuracy (98.25%), precision (98.09%), recall (98.15%) and F1-score (98.19%), using a real-time feedback and adaptive learning pathways. The findings demonstrate the effectiveness of combining advanced emotion recognition techniques with language learning, improving both learner engagement and language acquisition outcomes.