Intelligent system design to monitor organizational climate at a university using TinyML and emotion detection

This paper discusses the development of an automatic facial emotion recognition system employing a MiniXception architecture optimized for low-power devices like the Raspberry Pi 5. The primary aim is to analyze faculty emotions at a university and provide insights into the organizational climate. Utilizing facial images from a USB camera, the system classifies seven basic emotions - anger, disgust, fear, happiness, sadness, surprise, and neutrality - with an accuracy of 93.57%. The system incorporates image preprocessing and deep learning techniques for real-time results, displaying emotional feedback through text, emojis, and motivational audio messages. Detected emotions are logged with timestamps for future analysis. Testing with faculty members confirmed the system's accuracy and usability, suggesting its potential as a non-intrusive tool for continuous emotional monitoring and enhancing institutional well-being.