Current students


Section: Computer Science and Engineering

Major Research topic:
Semantic Human Behavior Recognition From Mobile Robots

With the evolution of mobile robots and their entrance to people's lives, the need for semantic effective human behavior recognition from mobile robots has become essential. The research project is to design the robust deep learning algorithm enabling human activity recognition based on semantic understanding from a mobile robot. The core technique is developing the robust on-board deep learning algorithm which elaborates the extracted skeletal joints of persons or objects from standard RGB cameras equipped on
mobile robots in real-time. With the help of semantic factors, the robots can recognize much easier different levels of human beings' activities than the traditional recognition algorithms: the scene level; the human-object interaction level; the human-human interaction level; the group people activity level and the group people interaction level. We will do the experiments basing on not only our new collected specific dataset, but also the existing widely known public datasets. Once having gotten the good performance, we will
also apply it in the real-life scenarios. The mature robot products will be applied in many public security applications like shopping mall, airport and subway.