Design and implementation of a facial expression recognition system based on a CNN model

Facial expression recognition is a critical task in human-computer interaction, with applications spanning emotional analysis, behavioral studies, and AI-driven interfaces. This study focuses on designing a facial expression recognition system based on convolutional neural networks (CNN) to classify facial images into six basic emotions: anger, fear, happiness, neutrality, sadness, and surprise. The study investigates how to optimize the CNN model architecture and employ data augmentation techniques to improve classification accuracy. Additionally, it explores methods to address the feature ambiguity between emotion categories and enhance the model's robustness under complex lighting and occlusion conditions. The system achieves an accuracy over 66% for only 150 epochs on the test dataset, demonstrating its effectiveness in recognizing facial emotions. The test results indicate that the recognition accuracy for five out of the six emotions is relatively high.