Abstract: A method of providing structured data to be used in a front end viewing device includes building a plurality of layers with different levels of detail (LODs). The method utilizes a point cloud, and selects points into the LODs on random basis. The data sets associated with the LODs are complementary, so that the data may be combined in order to achieve higher resolution. The method maintains or reduces the size of the resulting LOD-structured file(s) in comparison with the original file(s) of the cloud.

Abstract: A method of forming and using an efficient image data structure involves providing a first model comprising three-dimensional spatial information for use in being displayed on a display device. Objects within the first model having sub-models are classified to form groups of objects, objects within a same group for being processed with a substantially similar process. A class LOD structure is constructed for each group, the class LOD structure built using the substantially similar process for the class and the LOD structure for representing each object within the class for supporting rendering thereof. A rendering strategy is defined for each Class, the rendering strategy for rendering each object within the class based on data within the LOD structure and in dependence upon a gaze direction of a spectator and a location of the spectator. The class LODs are merged to form an LOD structure for the first model.