Design a Human-Robot Interaction Framework to Detect Household Objects

In human-robot interaction scenarios, the ability to identify a single object from multiple objects is an important task for service robots. Although there has been recent progress in this area, it remains difficult for autonomous vision systems to recognize objects in natural conditions. The service robot should detect a particular object according to the user's demand. This paper describes a human robot interaction framework to detect a particular household object from multiple objects through text-based interaction. Haar Cascade Classifiers is used to detect objects and developed a user friendly interface for human-system interaction. The propose framework use color, size, or position information to distinguish the user requested object in multi object scenarios. Evaluation results shows that the system is quite effective to detect the target household object from multiple objects in real time.