An optimized cloud-based mobile augmented reality system for marketing and advertising

点击翻译按钮获取中文摘要

Augmented Reality (AR) is a live, direct or indirect, view of a physical, real-world environment whose elements are augmented by computer-generated contents such as text, sound, video, or animation. AR marketing and advertising applications can add value to the product itself, either in the form of an enhanced product or a more engaging viewer experience. Collaborating with SimentIT, Inc., we will develop new technology in the cloud-based mobile AR system for marketing and advertising. Specifically, we will investigate three sub-projects: 1) object recognition, in which we first utilize the corner features and superpixels to create a semantic bag-of-words (BoW) model, and then apply a hierarchical Conditional Random Field (CRF) model to perform the object recognition; 2) tracking, in which we first propose a real-time tracking scheme based on Simultaneous Localization And Mapping (SLAM), and then improve the tracking performance by using Kernel Entropy Component Analysis (KECA) based multimodal information fusion; 3) game streaming, in which we propose a prioritized delivery scheme to maximize the scene quality by optimally allocating the bit rate for each 3D object, and a scalable Peer-to-Peer (P2P) geometry streaming scheme to provide different playback qualities for heterogeneous users. The proposed methods are expected to significantly improve the Quality of Experience (QoE) for mobile AR applications. Moreover, the proposed research will help to position Canada as a leader in multimedia technologies in the 21st century. The resultant technology transfer to Canadian industry will strengthen Canada's global competitiveness and create positive impacts to Canadian economy and society.