Abstract: The invention automatically produces a rendered image version of an original image that accurately represents perceptually distinguishable objects, surfaces and edges in the original image. The original image is first converted to a unique data structure representation called an image structure map (ISM) that accurately, compactly and explicitly represents the geometry, topology and signal properties of perceptually distinguishable regions in the original image. The ISM is a dynamic partition of the image induced by a set of image region boundaries, and includes, for each region in the original image, a region data item that indicates the region boundary data items that form the region and a signal property descriptor indicating the value of signals computed for image locations in the region. An automatic segmentation operation determines image region boundaries occurring in the original image by analyzing discontinuities in signal values of original image locations.

Abstract: A data structure representation of an original image called an image structure map (ISM) accurately, compactly and explicitly represents the geometry, topology and signal properties of regions in an original image. In mathematical terms, the ISM is a partition of the original image induced by a set of image region boundaries. The ISM data structure includes, for each region in the original image, a region data item that indicates the region boundary data items that form the region and a signal property descriptor indicating the value of signals computed for or assigned to the region. Region boundary data items represent image region boundaries that are determined to occur in the original image as a result of analyzing signal properties of original image locations.

Abstract: The present invention produces a data structure that indicates a partition of a given input set of line segments in a plane using a technique that is mathematically robust, canonical and dynamic. The technique is robust because it assumes a finite precision model of computer arithmetic and rounds the endpoints and intersections of all line segments to representable points in a way that is globally topologically consistent with the input set of line segments and that keeps the position of each rounded line segment close to the position of the input segment. The technique is canonical because the output partition produced is a function of the set of segments currently present only, and not of the history of insertion and deletions. This canonical aspect of the technique is facilitated by storing the input unrounded line segments in the partition data structure so that they are associated with their rounded fragments.

Abstract: A data structure representation of an image called an image structure map (ISM) accurately and explicitly represents the geometry, topology and signal properties of regions in an original image and allows for efficient and accurate spatial indexing of the image and those regions. The ISM may serve as the basis for an image interaction system in which a user interacts with a rendered view of an original image that is produced from the ISM data structure. The rendered view of the original image represents the geometry and topology of the original image as represented in the ISM. The user interacts with the rendered view to make alterations to, or to query, the structures in the ISM. The user may also modify the ISM by interacting directly with a displayed view of the original image. The user may further invoke automatic operations to be performed using the original image data structure that produce image region data, such as signal property descriptors and boundary data, that may be used to modify the ISM.