Facial Negative Emotion Recognition of Special Employees Based on Machine Vision

With the advancement of computer vision technology, emotion recognition has become a research hotspot in recent years due to its many applications, including mental state assessment, mental illness detection, human behavior understanding, and psychological analysis. For personnel working long-term in enclosed spaces such as military bases or cabins, more subtle negative emotions may lead to unsafe behaviors, resulting in accidents. To address this issue, this paper employs an optimized MTCNN face detection algorithm and an improved Video Swin Transformer model network to detect negative emotions in such specialized personnel based on facial features. The test results show that the overall accuracy exceeds 80%, outperforming existing emotion classifiers. This greatly reduces the likelihood of safety accidents caused by emotional state-induced operational errors, providing new insights and inspiration for future research.