Shell Representation and Compression Conscious Manipulation for Three Dimensional Graphical Datasets

• Authors: Martin J. Turner
• Venue: TPCG '03 Proceedings of the Theory and Practice of Computer Graphics 2003
• Year: 2003

Although there has been a large number of standardsfor two-dimension images, there has not been as extensivea study into representing three-dimensional datasets. Presentedhere is a test from an extended system to cater forinitially autostereo sequence datasets, although still suitablefor all types of three-dimensional datasets. Extensionsto do three dimensional manipulation operations are definedthat are both computationally efficient when appliedto the coded form and are compression conscious in thatthey maintain compression performance.Autostereoscopic displays have been developed whichcreate a three-dimensional image without the need for specialglasses or headsets. They work by displaying multipleimages each viewable over a very narrow field of view. Thishas the advantage of being able to look around objects.Increasing the number of images per view, increases thebandwidth required. To aid stereo disparity in the humanvisual system the important features to retain in stereo setsare potentially different from standard two-dimensional images.Edge information and highlights have been shown tobe of paramount importance for the human visual system toaccurately perceive depth.The theme of coding three-dimensional regions as separateentities allows for a potential high compression ratioand subsequent transmission rate speed-up. A controlledmerging scheme of regions is also presented which, beingdesigned to keep all highlights and avoid defocusing of theimage, gives a corresponding compression ratio increasewhile retaining the majority of the stereo cues.