Research on Facial Emotion Analysis Based on Multi-layer Convolutional Neural Network Method

This study aims to improve existing facial emotion recognition technology through a convolutional neural network (CNN) model. With the widespread demand for facial emotion analysis in multiple fields, this study proposes a multi-classification CNN model that can improve the accuracy and real-time performance of emotion recognition by optimizing the convolutional layer structure and improving data processing methods. This study used a seven-category emotion recognition task and conducted model validation based on multiple public datasets. Experimental results show that the improved CNN model shows high accuracy up to 65% when processing multiple emotion classifications, especially in the recognition of emotions such as "happiness" and "surprise" with significant improvements. Nonetheless, the model still needs to be further optimized when dealing with "fear" and "sadness" emotions. This research provides theoretical support and technical reference for the practical application of facial emotion recognition and promotes the development of intelligent systems in fields such as human-computer interaction.