Granted Patent B2

US 12,682,327 · App. 19/399,992 · Granted Jul 14, 2026

Inventors: Marlen L. Foster (Fairfield, CA); Connie K. Yung (San Francisco, CA); Vijay Narayanan (San Jose, CA); Raj M. Bharadwaj (Plymouth, MN); Soumitri Naga Kolavennu (Plymouth, MN); Jessica G. Winberg (Prescott, WI); Balasubramanian Narayanan (Minneapolis, MN); Hima Rama Subrahmanyam Vishnubhotla (Suwanee, GA); Jayalakshmi Mangalagiri (Lakeville, MN); Ali Marjani (Cary, NC); Teja Dade (Omaha, NE); Michael Starace (Oakland, CA); David Robinson (Burnsville, MN)

Assignee: U.S. Bank National Association

G06Q20/0425G06Q20/108G06Q20/3223G06Q20/326G06Q20/4016G06Q20/42G06Q40/02G06V30/19013G06V30/19093G06V30/1916G06V30/2253

Abstract

Claims (108)

1 . A method comprising:

receiving an image having a region of interest;

performing a binary conversion on the image to produce a binarized image;

generating contours for the region of interest in the binarized image;

applying a convex hull operation to the generated contours for the binarized image;

obtaining a set of corners from the generated contours following the application of the convex hull operation;

evaluating the set of corners;

determining that the set of corners passes the evaluation;

evaluating a quality of the image;

determining that the image passes the evaluation of quality; and

providing the image for further processing after determining that the set of corners passes the evaluation and that the image passes the evaluation of quality.

2 . The method of claim 1 , wherein evaluating the set of corners includes performing at least one operation from the corner evaluation group consisting of:

evaluating absolute size defined by the corners;

evaluating relative size defined by the corners;

evaluating whether a region of interest is in frame based on the corners;

evaluating corner angles of the corners;

validating consistency based on the corners; and

validating an aspect ratio based on the corners.

3 . The method of claim 2 , wherein evaluating the set of corners includes performing at least three operations from the corner evaluation group.

4 . The method of claim 2 , wherein evaluating the set of corners includes performing at least five operations from the corner evaluation group.

5 . The method of claim 2 ,

wherein evaluating absolute size defined by the corners includes deriving a height from at least two of the corners and a length from at least two of the corners and comparing the height and length to respective thresholds;

wherein evaluating relative size defined by the corners includes calculating an area of the image bounded by the contours relative the overall size of the image;

wherein evaluating whether a region of interest is in frame based on the corners includes determining whether the corners lie within a threshold distance away from an edge of the image;

wherein evaluating corner angles of the corners includes determining whether angles of the corners fall within a predetermined acceptable range;

wherein validating consistency based on the corners includes:

determining that a left height of a left side and a right height of a right side are within a certain threshold amount or percentage of each other; and

determining that a top length of a top side and a bottom length of a bottom side are within a certain threshold amount or percentage of each other; and

wherein validating an aspect ratio based on the corners includes deriving a height from at least two of the corners and a length from at least two of the corners and comparing a ratio based on the height and length to an acceptable ratio.

6 . The method of claim 1 , wherein evaluating the quality of the image includes performing at least one operation from a quality evaluation group consisting of:

a noise detection operation;

a minimal background validation operation;

a brightness validation operation;

a contrast validation operation;

a text contrast validation operation;

an image focus validation operation; and

a glare detection operation.

7 . The method of claim 6 , wherein evaluating the quality includes performing at least three operations from the quality evaluation group.

8 . The method of claim 6 , wherein evaluating the quality includes performing at least five operations from the quality evaluation group.

9 . The method of claim 6 ,

wherein the noise detection operation includes:

comparing the image to a blurred version of the image to form a comparison;

measuring a variance of pixel intensities based on the comparison; and

flagging the image as noisy responsive to the variance being below a predetermined threshold;

wherein the minimal background validation operation includes:

detecting a main rectangle in the image;

determining a ratio of the area of the main rectangle to an overall area of the image; and

flagging the image responsive to the ratio exceeding a threshold;

wherein the brightness validation operation includes:

creating a grayscale version of the image;

determining an average brightness of the grayscale version;

flagging the image responsive to the average brightness not falling within an acceptable brightness range;

wherein the contrast validation operation includes:

creating a grayscale version of the image;

determining an average contrast of the grayscale version;

flagging the image responsive to the average contrast not falling within an acceptable contrast range;

wherein the text contrast validation operation includes:

calculating a text region average pixel intensity of a text region in the image;

calculating a non-text region average pixel intensity of a non-text region in the image;

flagging the image if a ratio between the text region pixel intensity and the non-text region average pixel intensity does not satisfy a threshold;

wherein the image focus validation operation includes:

applying a filter to a grayscale version of the image;

measuring a variance based on the filtered grayscale version; and

flagging the image responsive to the variance failing to satisfy a threshold; and

wherein the glare detection operation includes:

calculating an area of regions having significant brightness in a grayscale version of the image;

comparing the area of regions having significant brightness with a total image area of the image to form a ratio; and

flagging the image responsive to the ratio exceeding or failing to satisfy a threshold.

10 . The method of claim 1 , further comprising:

performing one or more image conversion or cleanup steps with respect to the image.

11 . The method of claim 1 , further comprising:

applying one or more morphological modifications to the binarized image selected from the group consisting of: light erosion, heavy erosion, inversion with light erosion, and inversion with heavy erosion.

12 . The method of claim 1 , further comprising:

applying a correction selected from the group consisting of: a position normalizing operation; a proportion normalizing operation; a skewing operation; a deskewing operation; a rotation operation; a scaling operation; a stretching operation; and a moving operation.

13 . The method of claim 1 , wherein determining that the image passes the evaluation of quality includes determining that the evaluation of quality did not flag the image.

14 . The method of claim 1 , wherein providing the image for further processing includes applying an optical character recognition technique to the image.

15 . The method of claim 1 , wherein the method is performed by one or more processors of a mobile device having a camera.

16 . The method of claim 1 , wherein the region of interest is a check within the image.

17 . The method of claim 16 ,

wherein generating the contours for the region of interest includes incorrectly generating a convexity within the check; and

wherein applying the convex hull operation results in correcting the convexity.

18 . The method of claim 1 , further comprising:

capturing the image using a camera of a device; and

wherein the device performs the method.

19 . A set of one or more non-transitory computer-readable media having instructions thereon that when executed by a set of one or more processors, cause the set of one or more processors to:

receive an image having a region of interest;

perform a binary conversion on the image to produce a binarized image;

generate contours for the region of interest in the binarized image;

apply a convex hull operation to the generated contours for the binarized image;

obtain a set of corners from the generated contours following the application of the convex hull operation;

evaluate the set of corners;

determine that the set of corners passes the evaluation;

evaluate a quality of the image;

determine that the image passes the evaluation of quality; and

provide the image for further processing after determining that the set of corners passes the evaluation and that the image passes the evaluation of quality.

20 . An apparatus comprising:

a set of one or more processors; and

a set of one or more non-transitory computer-readable media having instructions thereon that when executed by the set of one or more processors, cause the set of one or more processors to:

receive an image having a region of interest;

perform a binary conversion on the image to produce a binarized image;

generate contours for the region of interest in the binarized image;

apply a convex hull operation to the generated contours for the binarized image;

obtain a set of corners from the generated contours following the application of the convex hull operation;

evaluate the set of corners;

determine that the set of corners passes the evaluation;

evaluate a quality of the image;

determine that the image passes the evaluation of quality; and

provide the image for further processing after determining that the set of corners passes the evaluation and that the image passes the evaluation of quality.

Assignments (1)

Continuity (6)

References Cited (270)