METHOD AND APPARATUS OF EXTRACTING PARTICULAR INFORMATION FROM STANDARD CARD

• US 20150317529A1
• Filed: 05/04/2015
• Published: 11/05/2015
• Est. Priority Date: 05/04/2014
• Status: Active Grant

• Current Assignee: Banma Zhixing Network (Hongkong) Co., Limited
• Original Assignee: Alibaba Group Holding Ltd.
• Inventors: Zhou, Shiqi, Chu, Refeng
• Granted Patent: US 9,665,787 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/1444 Selective acquisition, locating or processing of specific regions, e.g. highlighted text, fiducial marks or predetermined fields
• CPC Class Codes: G06V 30/1607 Correcting image deformation, e.g. trapezoidal deformation caused by perspective
• CPC Class Codes: G06V 30/224 of printed characters having additional code marks or containing code marks

Abstract

A method of extracting particular information in a standard card is disclosed herein. The method includes: acquiring a card image of a standard card having particular information to be extracted; identifying an image region containing the particular information in the card image; and extracting and outputting the image region as an independent image. Thus, an image related to the part of the particular information only can be obtained from the standard card conveniently, quickly and accurately, thereby improving the working efficiency. In addition, the present disclosure further provides an apparatus of extracting particular information in a standard card and a method of inputting particular information of a standard card in a mobile terminal.

Claims

1. A method implemented by one or more computing devices, the method comprising:

acquiring a card image of a standard card containing particular information to be extracted;

identifying an image region containing the particular information in the card image; and

extracting and outputting the image region as an independent image.

2. The method of claim 1, wherein acquiring the card image of the standard card includes:

acquiring a source image of the card image of the standard card containing the particular information to be extracted;

defining a region of the card image in the source image; and

extracting image information of the region of the card image in the source image as the card image of the standard card containing the particular information to be extracted.

3. The method of claim 2, wherein defining the region of the card image in the source image includes:

extracting image elements that are straight line segments and have a length greater than a set threshold in the source image;

combining the extracted straight line segments to determine combinations of line segments that are capable of forming a quadrilateral;

extending lengths of a plurality of line segments to determine positions of intersection points of the plurality of line segments in each combination of the combinations of line segments that are capable of forming the quadrilateral;

selecting one or more combinations of line segments satisfying a first condition, the first condition comprising a condition that the intersection points of the plurality of line segments are located within a range of the source image;

determining whether only one combination of line segments satisfies the first condition; and

in response to determining that only one combination of line segments satisfies the first condition, defining a region framed by the only one combination of line segments as the region of the card image in the source image.

4. The method of claim 3, wherein:

in response to determining that multiple combination of line segments satisfy the first condition, the method further comprises;

extracting the multiple combinations of line segments that satisfy the first condition;

screening the multiple combinations of line segments that satisfy the first condition based at least in part on a second condition, the screening comprising;

determining, in a respective combination of line segments of the multiple combinations of line segments, whether intersection points of line segments are on respective extension lines of the line segments, and taking the respective combination of line segments as a candidate combination of line segments if affirmative;

determining whether only one candidate combination of line segments exists; and

in response to determining that only one candidate combination of line segments exists, defining a region framed by the only one candidate combination of line segments as the region of the card image in the source image.

5. The method of claim 4, wherein a determination result is negative in the determining of whether only one candidate combination of line segments exists, screening candidate combinations of line segments by:

acquiring intersection points of line segments in a combination of line segments, the combination of line segments including four intersection points;

obtaining image coordinates of the four intersection points, mapping the image coordinates of the four intersection points to image coordinates of four points of a standard card template image to obtain a homography matrix, forming X and Y axes from two line segments using the homography matrix and using each intersection point in the four intersection points as an original point, computing unit vectors of X′ and

Y′

axes corresponding to the X and Y axes in the image after a homography matrix transformation, and computing a degree of an included angle between two unit vectors;

determining whether the acquired degree of the included angle between the two unit vectors in a combination of line segments is different from ninety degrees by an amount less than a preset threshold; and

defining a region framed by the combination of line segments as the region of the card image in the source image if affirmative.

6. The method of claim 1, wherein identifying the image region containing the particular information in the card image includes:

acquiring relative position parameters that reflect a relative position of the particular information in the standard card based on parameters of a standard card template image;

performing a homography matrix transformation on the card image based on a shape of the standard card template image to acquire a transformed card image; and

clipping an image region in conformity with the relative position parameters in the transformed card image based on the relative position parameters of the particular information in the standard card template image, the clipped image region being the image region containing the particular information.

7. The method of claim 1, wherein identifying the image region containing the particular information in the card image includes:

performing a homography matrix transformation on the card image based on a standard card template image to acquire a transformed card image;

filtering the standard card template image to acquire a filtered standard card template image;

acquiring a first pixel value summation distribution diagram of the standard card template image in a recording direction of the particular information to be extracted based on the recording direction;

filtering the transformed card image to acquire a filtered card image;

performing a pixel value summation computation in the recording direction of the particular information in the filtered card image based on the recording direction to acquire a second pixel value summation distribution diagram of the card image in the recording direction; and

comparing the second pixel value summation distribution diagram with the first pixel value summation distribution diagram, locating, in the second pixel value summation distribution diagram, a position having a same pixel value summation feature in the recording direction of the particular information in the first pixel value summation distribution diagram, and identifying thereof as the image region containing the particular information in the card image.

8. The method of claim 7, wherein the standard card comprises a bank card, the particular information comprises a card number and a peak region of the second pixel value summation distribution diagram is used as the image region containing the particular information.

9. The method of claim 8, wherein the card number is produced by embossing.

10. The method of claim 7, wherein relative position parameters of the particular information in the standard card template image are simultaneously referenced when locating the position having the same pixel value summation feature in the recording direction of the particular information in the pixel value summation distribution diagram of the standard card template image.

11. The method of claim 1, wherein identifying the image region containing the particular information in the card image includes:

performing a homography matrix transformation on the card image based on a shape of a standard card template image to acquire a transformed card image;

determining a size occupied by the particular information of the standard card template image in the standard card template image based on parameters of the standard card template image;

filtering the transformed card image to acquire a filtered card image;

performing a pixel value summation computation in a recording direction of the particular information in the filtered card image based on the recording direction to acquire a pixel value summation distribution diagram of the card image in the recording direction;

determining a range region of the particular information in the pixel value summation distribution diagram according to the size occupied by the particular information on the standard card template image; and

mapping a position of the particular information in the pixel value summation distribution diagram to the transformed card image, and extracting the image region containing the particular information of the transformed card image from the range region.

12. The method of claim 11, wherein determining the range region of the particular information in the pixel value summation distribution diagram according to the size occupied by the particular information in the standard card template image includes:

multiplying a size of the particular information by a range coefficient to compute a position range of the particular information in the pixel summation distribution diagram;

selecting a peak value within the position range of the pixel summation distribution diagram; and

selecting the range region in the pixel summation distribution diagram according to the size of the particular information using the peak value as a center, and mapping the range region to the transformed card image.

13. The method d of claim 1, wherein after extracting and outputting the image region as the independent image, an OCR technology is used to perform information identification on the outputted image region to acquire a textual form of recorded information.

14. An apparatus comprising:

one or more processors;

memory;

an acquisition unit stored in the memory and executable by the one or more processors to acquire a card image of a standard card containing particular information to be extracted;

an identification unit stored in the memory and executable by the one or more processors to identify an image region containing the particular information in the card image acquired by the acquisition unit; and

an output unit stored in the memory and executable by the one or more processors to output the image region identified by the identification unit as an independent image.

15. The apparatus of claim 14, wherein the acquisition unit includes:

a source image acquisition unit to acquire a source image of the card image of the standard card containing the particular information to be extracted;

a defining unit to define a region of the card image in the source image acquired by the source image acquisition unit; and

an extraction unit to extract image information from the region defined by the defining unit as the card image of the standard card containing the particular information to be extracted.

16. The apparatus of claim 15, wherein the defining unit includes:

an image element extraction unit to extract image elements which are straight line segments and have a length greater than a set threshold in the source image;

a combination unit to combine the straight line segments extracted by the image element extraction unit and determine combinations of line segments that are capable of forming a quadrilateral;

an intersection point determination unit to extend lengths of four line segments in a combination of the combinations of line segments of the combination unit, and determine positions of intersection points of the line segments;

a first condition selection unit to take the intersection points of the line segments determined by the intersection point determination unit being within a range of the source image as a first condition, and select a combination of line segments that satisfies the condition;

a first condition determination unit to determine whether only one combination of line segments selected by the first condition selection unit satisfies the condition, and initiate a card image determination unit if affirmative; and

the card image determination unit to define a region framed by the combination of line segments as the region of the card image in the source image.

17. The apparatus of claim 16, wherein:

if a determination result of the first condition determination unit is negative, a first condition extraction unit is initiated, and wherein

the first condition extraction unit extracts all combinations of line segments that are selected by the first condition selection unit and satisfy the first condition;

a second condition first judgment unit screens the combinations of line segments that are extracted by the first condition extraction unit according to a second condition by;

determining whether intersection points of line segments are on respective extension lines of the line segments in a combination of line segments, and treating the line segments as a candidate combination of line segments if affirmative; and

a second condition second determination unit determines whether only one combination of line segments that is acquired by the second condition first determination unit satisfies the second condition, and defines a region framed by the combination of line segments as the region of the card image in the source image if affirmative.

18. The apparatus of claim 17, wherein if a determination result is negative in the determination of whether only one combination of line segments exists by the second condition second determination unit, combinations of line segments that satisfy the second condition are screened by:

an intersection point acquisition unit to acquire intersection points of line segments in a combination of line segments from the second condition second determination unit, the combination of line segments including four intersection points;

a computation unit to obtain image coordinates of the four intersection points, and map image coordinates of four points of the standard card template image to the image coordinates of the four intersection points to obtain a homography matrix, compute X and Y axes that are formed from two line segments using the homography matrix and using each intersection point of the four intersection points as an original point, compute unit vectors of X′ and

Y′

axes corresponding to the X and Y axes in the image after a homography matrix transformation, and compute an included angle between the unit vectors; and

a threshold determination unit to determine whether the acquired included angle between the two unit vectors in a combination of line segments is different from ninety degrees by an amount less than a preset threshold, and define a region framed by the combination of line segments as the region of the card image in the source image if affirmative.

19. The apparatus of claim 14, wherein the identification unit comprises:

a parameter acquisition unit to acquire relative position parameters which reflect a relative position of the particular information in the standard card according to parameters of a standard card template image;

a transformation unit to perform a homography matrix transformation according to the card image extracted by the acquisition unit based on a shape of the standard card template image to acquire a transformed card image; and

a clipping unit to clip an image region in conformity with the relative position parameters in the transformed card image according to the relative position parameters of the particular information in the standard card template that are acquired by the parameter acquisition unit, the clipped image region being the image region containing the particular information.

20. One or more computer readable media storing executable instructions that, when executed by one or more processors, cause the one or more processors to perform acts comprising:

storing a standard card template image in a mobile terminal;

extracting an image region containing particular information from a standard card based at least in part on relevant parameters, a size and/or a shape of the standard card template image; and

outputting the extracted image region containing the particular information to a corresponding interface of the mobile terminal.