System and method for compressing image files while preserving visually significant aspects
OwnerGregory Michael Burkeinventor
Life of this patent
Estimating the term…
Abstract
The present invention concerns a way of compressing an initial data structure that defines an image as an array of pixel elements. The method includes generating a new data structure that defines contours at least some of which are contours having a constant color value to within a tolerance.
Claims49
A method of transmuting (i) image data, meaning data representing an image, in the form of pixels having both positional point coordinates and color values, into (ii) another form of data representing the same image, for purposes including image display, the method comprising:
- from the image data, defining contour data representing the image in terms of points on contours connecting multiple points each of which points has the same color value to within a predetermined tolerance, the total number of different contours being determined as the minimum required to provide an interpolated fit between contours that will keep error in any visual representation of the image from the contour data below a predetermined threshold error;
- and
- displaying the image from the contour data;
- wherein by action of the defining the image data in the form of pixels is transmuted into a new from of data, called contour data, having the form of contours each of which connects multiple points having the same color value to within the predetermined tolerance;
- wherein insofar as some image points of identical color values need not both appear on any contours, then the contour data is first-abbreviated, or first-compressed, from the image data from which it is derived.
A method of compressing an initial data structure, wherein the initial data structure defines an image as an array of pixel elements, the method comprising:
- providing an error threshold;
- generating a new data structure defining a plurality of contours, wherein for at least some of the contours each contour has a color value that is constant to within a tolerance;
- and
- adding information to the new data structure that defines additional contours until an interpolative error between the contours is below the error threshold.
A method of compressing a first data structure wherein the initial data structure defines an image as an array of pixel elements, the method comprising:
- generating a second data structure defining a plurality of contours wherein generating the data structure for each of at least some of the plurality of contours includes:
- identifying individual points along the contour, the points having a color value that is constant to within a tolerance;
- fitting at least one curve to the contour;
- and
- storing control points that define the at least one curve.
A method of compressing a first data structure wherein the initial data structure defines an image having both positional point coordinates and color values as an array of pixel elements, the method comprising:
- first generating a new data structure that defines from the initial data structure a plurality of primitives, each of at least some of the primitives bounded by a plurality of contour curves connecting multiple points each of which points has the same color value to within a predetermined tolerance, the total number of different contour curves being determined as the minimum required to provide an interpolated fit between curves that will keep error in any visual representation of the image from the new data structure below a predetermined threshold error,
- second generating at least one cross connect that couples one contour curve to another contour curve.
A computer program stored in a computer-readable medium, for processing a data structure representing an image, in the form of pixels having both positional point coordinates and color values
- wherein the collective pixels define a plurality of contours wherein the contours define some regions of substantially constant color and some regions of varying color;
- and
- wherein the number of contours are determined as the minimum required to provide an interpolated fit between contours that will keep error in any visual representation of the image from the pixel data below a predetermined threshold error.
A method of generating an image comprising:
- providing a data structure defining a plurality of contours wherein the contours define some regions of substantially constant color and some regions of varying color, the total number of different contours being determined as the minimum required to provide an interpolated fit between contours that will keen error in any visual representation of the image from the data structure below a predetermined threshold error;
- and
- defining connections between the contours to form polygons.
A method of generating an image comprising:
- providing a data structure having points with the same color value within a predetermined tolerance, the points defining a plurality of contours wherein the contours define some regions of substantially constant color and some regions of varying color, and wherein the total number of different contours are determined as the minimum required to provide an interpolated fit between contours that will keep error in any visual representation of the image from the contour data below a predetermined threshold error;
- and
- adding or removing points from the data structure based on an image magnification factor.
Dependencies are inferred from the claim text.