Facial Expression Recognition in Museum Settings Using an Enhanced Residual Networks

Enhancing visitor interaction is crucial in the context of digital exhibitions in museums. This paper proposes a facial expression recognition method based on ResNet18, augmented with attention mechanisms and spatial pyramid pooling, to improve accuracy under complex lighting and background conditions. Utilizing ResNet18 for facial feature extraction, combined with Convolutional Block Attention Module (CBAM) and Spatial Pyramid Pooling (SPP), the method achieves efficient recognition of six basic expressions: happiness, fear, sadness, anger, surprise, and disgust. Experimental results on the FER2013 dataset show a peak recognition accuracy of 73.22%, demonstrating significant improvement in expression recognition.