Improving the Reader’s Attention and Focus through an AI-Driven Interactive and User-Aware Virtual Assistant for Handheld Devices

Abstract

This paper describes the design and development of an AI-driven, interactive and user-aware virtual assistant aimed at helping users to focus their attention on reading or attending to other long-lasting visual tasks. The proposed approach uses computer vision and artificial intelligence to analyze the orientation of the head and the gaze of the user’s eyes to estimate the level of attention during the task, as well as administer effective and balanced stimuli to correct significant deviations. The stimuli are provided by a graphical character (i.e., the virtual assistant), which is able to emulate face expressions, generate spoken messages and produce deictic visual cues to better involve the user and establish an effective, natural and enjoyable experience. The described virtual assistant is based on a modular architecture that can be scaled to support a wide range of applications, from virtual and blended collaborative spaces to mobile devices. In particular, this paper focuses on an application designed to integrate seamlessly into tablets and e-book readers to provide its services in mobility and exactly when and where needed.