Accented visualization application in interactive manuals for technical training and support

Abstract

The capabilities of modern information and communication devices (tablets, glasses, helmets, etc.) that provide user interfaces with Augmented Reality (AR) are successfully used in education and maintenance support nowadays. The main benefits are concerned with the possibility to display the required information in the context of operations performed by the user. Technically it is implemented in the form of textual hints or visual 3D objects with highly realistic appearance and behavior. New features prompt new challenges: The user can be easily overloaded with lots of information, and some real objects that require user's attention can be overlapped by virtual ones in one scene. In order to solve this problem and improve the usability and effectiveness of AR devices application in maintenance and education it is proposed to build adaptive user interfaces that consider the context and focus of the user. From information technology perspective this approach forms a promising area of application of high-potential technologies including machine learning and computer vision. In this paper there is presented an example of AR implementation as a part of an interactive manual, which provides adaptation and personalization based on the concept of accented visualization. Identification of objects and their classification is implemented using the artificial neural network.