Data compression using multiple videos for photo-based augmented reality system

摘要

A photo-based augmented reality (Photo AR) system is an AR system that uses photographic images, instead of computer graphics of 3D models. The advantage of this system is that it can display realistic augmented views to users; however, it requires large disk-space volumes because it uses many images. The images can be compressed by storing in video formats; however, the rendering times of the conventional method may sometimes cause severe delays while displaying objects. In order to ease this problem, we propose a method where all photographic images are stored into multiple video files, extracted in the memory at startup, and processed in parallel by the graphics processing unit (GPU) to save the time required for seeking and decoding video files.