Emotion based Media Playback System using PPG Signal

The study involved identifying human emotions and integrates the identified emotion with the music system. The idea is to develop a complete product to utilize the detected emotion in a real-time application and also to achieve more accuracy and less memory. Human emotions are identified using physiological signals such as electrocardiography, electromyography, photoplethysmography, respiration, skin temperature, etc. Obtaining photoplethysmography (PPG) from the sensor is a simple, cost-effective, and non-invasive method. PPG sensors are capable of providing accurate heart-rate (HR) by detecting the variations in the blood flow. Signals are acquired using wearable technology from a personal whereas not compromising comfort and privacy. The attributes of Heart Rate Variability are analyzed to describe emotions, namely happy, calm, unhappy (sad), and fear using Machine learning technology. We deployed this recognized emotion to automate the music system associated with its emotion. To bring this, we built an Android app to communicate with the smart wearable utilized. Totally 150 members from both genders have participated. The accuracy of 91.81% is achieved. This emotion recognition system can be used in various fields like robotics, medicine, virtual reality, and gaming, advertising, education, automotive working conditions and safety, home appliances.