Back to results
US 7,254,275 B2Granted
Method and system for image compression using image symmetry
AssigneeSymbol Technologies, Inc.
InventorsAli Tabesh, Duanfeng He
Life of this patent
Estimating the term…
Abstract
An image compression system and method is described, which makes use of the symmetry found in faces and heads to perform the compression. The image is divided along the line of symmetry, and pairs of corresponding pixels on the two divided sides are determined. A weighted average and a weighted variance of the pixel values of the pairs is computed, and is used to encode the image. A transform such as the Karhunen-Loeve transform is used to compute the weighted averages and variances.
Claims29
A method for reducing an image representing a substantially symmetric subject, comprising the steps of:
- dividing the image into a first part and a second part, the second part being substantially symmetric to the first part;
- determining pairs of symmetrically corresponding pixels, a member of the pair in the first part having an X pixel value and a member of the pair in the second part having a Y pixel value;
- computing a weighted average value A of the pixel values of each one of the pairs;
- computing a weighted variance value B of the pixel values of each one of the pairs, wherein the A and B values are computed as a function of a fixed transformation matrix K;
- representing each of the pairs with the weighted average and the weighted variance values;
- and
- generating reduced image data including at least a plurality of A values and corresponding B values.
4 more elements
A method for reducing an image representing a substantially symmetric subject, comprising:
- dividing the image into a first part and a second part, the second part being substantially symmetric to the first part;
- determining pairs of symmetrically corresponding pixels, a member of the paire in the first part having an X pixel value a member of the pair in the second part having a Y pixel value;
- computing a weighted average value A of the pixel values of each one of the pairs;
- computing a weighted variance value B of the pixel values of each one of the pairs;
- representing each of the pairs with the weighted average and the weighted variance values;
- generating reduced image data including at least a plurality of the pairs;
- reconstituting the image by computing the X pixel value and the Y pixel value from the weighted average value A and the weighted variance value B;
- and
- compressing the reduced image data with an image compression algorithm
- wherein the reconstituting step further includes the substep of removing visual artifacts from the reconstituted image, and
- wherein the visual artifacts comprise a seam line between the first part and the second part.
8 more elements
A method for reducing an image representing a substantially symmetric sugject, comprising:
- dividing the image into a first part and a second part, the second part being substantially symmetric to the first part;
- determining pairs of symmetrically corresponding pixels, a member of the pair in the first part having an X pixel value and a member of the pair in the second part having a Y pixel value;
- computing a weighted average value A of the pixel values of each one of the pairs;
- couputing a weighted variance value B of the pixel values of each one of the pairs;
- representing each of the pairs with the weighted average and the vaighted variance values;
- generating reduced image data including at least a plurality of the pairs;
- reconstituting the image by computing the X pixel value ang the Y pixel value from the weighted average value A and the weighted variance value B;
- and
- compressing the reduced image data with an image compression algorithm
- wherein the reconstituting step further includes the substep of removing visual artifacts from the reconstituted image, and
- wherein the removing step includes the substeps of comparing the weighted variance value to a difference in weighted average values for adjacent pixels.
8 more elements
A system for processing an image substantially symmetric along a line of symmetry, comprising:
- a software module configured to process pixel values of pixels representing the image, the software module being adapted to
- divide the image substantially along the line of symmetry;
- determine pairs of corresponding pixels, each pixel on a side of the divided image;
- compute a weighted average and a weighted variance of each of the pairs as a function of a fixed transformation matrix;
- and
- generate reduced image data from at least the weighted average and the weighted variance of the pairs;
- and
- a storage media adapted for storing the reduced image data.
5 more elements
A system for processing an image substantially symmetric along a line of symmetry, comprising:
- a storage media reader adapted to acquire reduced image data formed of at least weighted average values and weighted variance values of pairs of corresponding pixels, the corresponding pixels being on opposite sides of the line of symmetry;
- and
- a software module configured to compute pixel values of each of the pair of pixels by applying a fixed inverse transform to the weighted average values and the weighted variance values, and to reconstitute the image from the computed pixel values.
A stand alone identification system, comprising:
- a barcode scanner adapted to acquire weighted average and weighted variance data from a 2D barcode;
- a microprocessor adapted to execute software configured to:
- process the weighted average and weighted variance data to generate pixel values for pairs of corresponding pixels by applying a fixed inverse transform to the weighted average and weighted variance data;
- determine location of the corresponding pixels in an image, each of the pixels of the pair being on a side of a line of symmetry of the image;
- and
- reconstituting the image using the computed pixel values;
- and
- a display module adapted to display at least the reconstituted image.
5 more elements
Dependencies are inferred from the claim text.