In today’s technology, human-machine interaction is becoming more common, and robots must be able to comprehend human actions and reactions. Human behavior can be better understood if technology can recognize person’s emotions, resulting in increased work efficiency. Text, audio, linguistic, and facial expressions can all be used to communicate emotions. Facial expressions are important in judging a person’s emotions. In this work, an approach for classifying human moods using face pictures is presented using deep learning technique. The suggested approach uses Haar-cascade face recognition, Active shape Model (ASM) data retrieval (26 face points collected), and Ada-boost algorithm to categorize five emotional responses: rage, hate, pleased, neutrality, and amazement. The proposed approach is unique in that it does the emotional expression in actual environments on the Raspberry Pi, with a real-time average accuracy of 94%. In social surroundings where sentiment identification is vital, the Raspberry Pi can recognize sentiments flexibly in live time when it has been joined to a robotic system.