Depth Image Generation for Autostereoscopic Monitors

There have been many attempts to generate video for multiple view points using minimal data in order to use autostereoscopic displays which require input from multiple view points. Our research aims to establish technology to generate images with depth ("depth images") from two cameras or two images. This paper defines requirements for clear, three-dimensional depth images, and proposes both real-time and offline approach methods to depth image generation which meets these requirements.