Adjusting perspective distortion of an image
AssigneeGradeable, Inc.
InventorDante Cassanego
Life of this patent
Estimating the term…
Abstract
The present disclosure concerns a method and system to accurately remove a three-dimensional distortion in an image of a document and convert the image into an accurate two dimensional image. A method for accurately deducing an image of a document to a precise document boundary is also disclosed. A portable computing device may employ the disclosed method and system. The methods involve a marker embedded in a document which provides three-dimensional positional information of a recording device with reference to the marker continuously in real time.
Claims13
A method of correcting a three-dimensional distortion in an image of a document comprising the steps of:
- embedding a marker in the document, the marker including a predetermined marker orientation;
- capturing a full image that includes the image of the document, a marker orientation appearing in the captured full image, wherein the marker when captured in the full image is virtually altered to appear as the marker orientation, the marker orientation being a three-dimensional virtual projection of the marker and appearing in place of the marker;
- identifying the marker orientation of the marker within the image of the document;
- comparing the marker orientation of the marker within the image of the document to the predetermined marker orientation, wherein the predetermined marker orientation is a desired marker orientation, representing the marker orientation without the three-dimensional distortion;
- determining a transformation to remove the three-dimensional distortion in the image of the document;
- and
- applying the transformation to the captured full image.
A method of determining a document boundary in an image of a document comprising the steps of:
- embedding a marker at a predetermined location in the document, the marker including a predetermined information relative to the document size and shape, and to the marker orientation, location, size and shape;
- capturing a full image that includes the image of the document, a marker orientation appearing in the captured full image, wherein the marker when captured in the full image is virtually altered to appear as the marker orientation, the marker orientation being a three-dimensional virtual projection of the marker and appearing in place of the marker;
- identifying the marker orientation of the marker within the image of the document;
- comparing the marker orientation of the marker within the image of the document to a predetermined marker orientation, wherein the predetermined marker orientation is a desired marker orientation, representing the marker orientation without a three-dimensional distortion;
- determining a transformation to remove the three-dimensional distortion from the image of the document;
- and
- identifying the document boundary within the captured full image on the basis of at least the transformation, the document size and shape, the marker location, size, and shape.
Dependencies are inferred from the claim text.