Machine Perception for Identification and Interaction in the Internet of Things

The most significant step from the Internet towards the Internet of Things (IoT) is to embrace a vast amount of objects in the global system. This first presents a nontrivial challenge to identification due to the limited applicability of Internet Protocol (IP) to these objects. They are therefore identified either via the augmentation of tiny devices such as radio frequency identification (RFID) or via natural feature identification where machine perception is indispensable. Secondly, the value of the IoT consists in the object-object and object-human interaction. Such interaction is considerably different from traditional human-computer interaction because of the large variety of objects and their surroundings. This paper presents the application of machine perception in both identification and interaction, which again are useful in finding information shadows that physical objects cast in the digital world.