Emotion recognition and conflict early warning technology is a relatively new branch of edtech; its principles are usually based on computer vision, speech emotion analysis and behavior pattern recognition, collecting classroom data through cameras, microphones or wearable devices, and then using algorithms to determine students' emotional states or potential conflict risks.