Detecting Facial Expressions and Recognition in a Mask Wear Person

In the field of research, the automatic detection and classification of facial expressions holds significant importance for understanding human emotions, essential to effective communication. However, the advent of COVID-19 introduced an unexpected challenge the widespread wearing of face masks, which conceal key facial cues. Addressing this, our proposed technology integrates low-light image enhancement, facial landmark detection via Deep Convolution Neural Networks (DCNN), and the efficiencies of transfer learning using pre-trained MobileNetV3 models. Impressively achieving a 72.3% accuracy on the AffectNet dataset, this innovation promises to transform sectors ranging from healthcare, where understanding patient emotions is vital, to education, retail, and even security. Beyond just enhancing human-machine interfaces, this system fortifies interpersonal communication in our masked era, ensuring that emotions are consistently communicated and understood.