US 11,893,816 B2 · Document detection in digital images

Abstract

Methods and systems are presented for detecting a boundary of a document within a digital image. Upon receiving an image, the image is converted into a binary image. One or more kernel-based transformations are performed on the binary image using a horizontal kernel and a vertical kernel. A plurality of edges are identified based on the one or more kernel-based transformations. The plurality of edges includes a plurality of horizontal edges and a plurality of vertical edges. Multiple quadrilaterals are constructed using different combinations of horizontal edges and vertical edges from the plurality of edges. A particular quadrilateral is selected from the multiple quadrilaterals based on how well the edges fit the perimeters of the quadrilaterals. The selected quadrilateral is used to define a boundary of the document within the digital image.

Claims20

Dependent claims

A system, comprising:

a non-transitory memory;
and
one or more hardware processors coupled with the non-transitory memory and configured to read instructions from the non-transitory memory to cause the system to perform operations comprising:
receiving an image comprising a digital representation of a document;
generating a first transformed image based on performing a first transformation to the image using a horizontal line kernel;
identifying a plurality of horizontal edges on the image based on the first transformed image;
generating a second transformed image based on performing a second transformation to the image using a vertical line kernel;
identifying a plurality of vertical edges on the image based on the second transformed image;
constructing a quadrilateral using two vertical edges from the plurality of vertical edges and two horizontal edges from the plurality of horizonal edges;
and
extracting a portion of the image within the quadrilateral as the digital representation of the document.

8 more elements

A method, comprising:

receiving, by one or more hardware processors, an image of a document;
generating, by the one or more hardware processors, a first transformed image based on performing a first transformation to the image using a horizontal line kernel;
generating, by the one or more hardware processors, a second transformed image based on performing a second transformation to the image using a vertical line kernel;
identifying, by the one or more hardware processors, a plurality of edges within the image based on the first transformed image and the second transformed image, wherein the identifying plurality of edges comprises identifying a plurality of horizontal edges within the image based on the first transformed image and identifying a plurality of vertical edges within the image based on the second transformed image;
constructing, by the one or more hardware processors, a plurality of quadrilaterals based on the plurality of edges, wherein each quadrilateral in the plurality of quadrilaterals is constructed using different subset of edges in the plurality of edges;
selecting, by the one or more hardware processors, from the plurality of quadrilaterals, a particular quadrilateral based on a metric representing how well edges used to construct each corresponding quadrilateral in the plurality of quadrilaterals match a perimeter of the corresponding quadrilateral;
and
extracting, by the one or more hardware processors, a portion of the image based on the particular quadrilateral.

5 more elements

Dependencies are inferred from the claim text.

Document detection in digital images

Life of this patent

Document detection in digital images

Life of this patent

Abstract

Claims20