Evaluation of feature-based object identification for augmented reality applications on mobile devices

With the technological advancement in mobile computing industry the field of electronic commerce has gone through a paradigm shift. Today's customers are more inclined towards using sophisticated mobile assistants, to help them in shopping, indoor navigation etc. rather than struggling on their own. In the literature it can be found that several attempts have been taken to address this cause. Majority of those attempts are based on marker based object identification. But current scale and the trends of the electronic commerce industry demand shopping assistants who are following markerless object identification methods due to limitations like the limited number of objects that can be identified. In order to support more usable electronic commerce applications, this paper focuses on the feasibility of deploying common and readily available feature tracking algorithms which can provide the object identification capability without having fiducial markers.