Method of pushing context-aware service of augmented reality by scene classification

This dissertation presents a method of pushing context-aware service based on scene classification to overcome the improper use of labels in augmented reality browsers caused by the interference with user's cognitive operations. User's scene is separated by using the proposed method on the basis of user's cognitive processes when searching and retrieving points of interest. The function module of three scenes is built by using the framework of four-layer context-aware services respectively, then the overall construction of the augmented reality system is completed by associating three scenario modules. Experimental results of comparison with the existing popular browsers show that the classification accuracy of the system constructed by using the proposed method is improved by 13% on average, and the user's average satisfaction with the prediction results of the system is increased by about 26%.