Method and apparatus for image classification and halftone detection
AssigneeElectronics for Imaging, Inc.
InventorsRon J. Karidi, Niklas Nordstrom
Life of this patent
Estimating the term…
Abstract
A method and apparatus for image classification includes a first embodiment, in which halftone detection is performed based on the size of a boundary set between a class of light pixels and a class of dark pixels, and further based upon image information contained within each single image plane (i.e within one color plane). This embodiment of the invention is based upon the distinctive property of images that halftone areas within the image have larger boundary sets than non-halftone areas within the image. A second, equally preferred embodiment of the invention provides a cross color difference correlation technique that is used to detect halftone pixels.
Claims32
A method for image classification and halftone detection, comprising:
- selecting a subject pixel;
- separating neighbor pixels about the subject pixel into light and dark classes in a first neighborhood;
- measuring a size of a boundary set between the light and dark classes in a second neighborhood;
- determining if the boundary size is less than a first threshold, wherein the subject pixel is not a halftone pixel;
- and
- determining if the boundary size is equal to or greater than the first threshold, wherein the subject pixel is a halftone pixel, and wherein descreening techniques may optionally be applied to the subject pixel.
A method for image classification and halftone detection, comprising the steps of:
- selecting a subject pixel;
- separating neighbor pixels about said subject pixel into light and dark classes in a first neighborhood;
- measuring the size of a boundary between said light and dark classes in a second neighborhood;
- determining if said boundary size is less than a first threshold, wherein said subject pixel is not a halftone pixel;
- determining if said boundary size is equal to or greater than said first threshold, wherein said subject pixel is a halftone pixel, and wherein descreening techniques may optionally be applied to said subject pixel;
- and
- applying a cross correlation technique to detect halftone pixels.
An apparatus for image classification and halftone detection, comprising:
- a first neighborhood consisting of:
- a subject pixel;
- and
- neighbor pixels about said subject pixel;
- wherein said neighbor pixels about said subject pixel are separated into light and dark classes in said first neighborhood;
- a second neighborhood, wherein the size of a boundary set between said light and dark classes in said second neighborhood is measured;
- means for determining if said boundary size is less than a first threshold, wherein said subject pixel is not a halftone pixel;
- means for determining if said boundary size is equal to or greater than said first threshold, wherein said subject pixel is a halftone pixel;
- and
- means for optionally applying descreening techniques to said subject pixel.
An apparatus for image classification and halftone detection, comprising:
- means for selecting a subject pixel;
- means for separating neighbor pixels about said subject pixel into light and dark classes in a first neighborhood;
- means for measuring the size of a boundary between said light and dark classes in a second neighborhood;
- means for determining if said boundary size is less than a first threshold, wherein said subject pixel is not a halftone pixel;
- means for determining if said boundary size is equal to or greater than said first threshold, wherein said subject pixel is a halftone pixel, and wherein descreening techniques may optionally be applied to said subject pixel;
- and
- means for applying a cross correlation technique to detect halftone pixels if said boundary size is not equal to or greater than said first threshold;
- means for calculating a variational norm within a single component window;
- means for calculating correlation factors for said subtext pixel;
- and
- means for comparing said correlation factors with a second threshold, wherein said subject pixel is a halftone pixel if at least one of said correlation factors is less than said second threshold.
Dependencies are inferred from the claim text.