Contactless Interaction System Based on Facial Expression Recognition for Humanoid Piano Robot
Ling Zhong,Wen Wang,Shiqiang Zhu,Shanshan Ji,Changhai Zha,Minhong Wan,Jason Gu
DOI: https://doi.org/10.1109/robio55434.2022.10011857
2022-01-01
Abstract:With the wide application of service robots in people's daily life, people are not only satisfied with robots accomplishing tasks independently but also hope that robots can maintain sustainable interaction with human beings. Especially in the field of music, robots need to be equipped with intelligent cognitive skills in decision-making, so that the audience can enjoy immersive appreciation and emotional resonance in the robots performance. Visual perception and natural language understanding are essential for robots to establish human-robot friendly relationships. To enrich the interactive ability of music robots, we design a contactless interaction system based on facial expression recognition, which consists of four modules: voice wake-up, face detection, facial expression recognition, and music mapping. When people wake up the music robot with specific words, the robot will customize appropriate music to perform according to the recognition results of the FER module. As a final step, the interaction system has been successfully applied to the piano humanoid robot independently developed by our team.