Method for determining coordinates of circumscribed rectangular frame of each character for use in optical character reader

• US 5,054,091 A
• Filed: 04/18/1990
• Issued: 10/01/1991
• Est. Priority Date: 04/18/1989
• Status: Expired due to Fees

• Current Assignee: Sharp Kabushiki Kaisha (Hon Hai Precision Industry Co., Ltd.)
• Original Assignee: Sharp Kabushiki Kaisha (Hon Hai Precision Industry Co., Ltd.)
• Inventors: Saiga, Hisashi, Tanaka, Hideaki, Morita, Toshiaki, Nakamura, Yasuhisa, Kitamura, Yoshihiro, Hirose, Hitoshi
• Primary Examiner(s): Boudreau, Leo H.
• Application Number:
  US07/510,895
• Time in Patent Office: 531 Days
• Field of Search: 382/9, 382/48, 382/18
• US Class Current: 382/177
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/148 Segmentation of character regions

Abstract

A method determines coordinates of a circumscribed rectangular frame for enclosing each character based on image data of character images in a recognition area. First of all, specified image data of a predetermined data unit in a transverse direction including at least one bit of black pixel are determined by scanning the image data. Thereafter, top and bottom coordinates of a circumscribed rectangular frame are determined respectively by searching respective image data located at the top and bottom positions among plural image data including at least one bit of black pixel located on the upper and lower sides of the specified image data, wherein respective bits of black pixel of plural image data are substantially continuously connected in a longitudinal direction, and then, leftmost and rightmost coordinates of the circumscribed rectangular frame are determined based on histogram data composed of either logical products or logical sums of respective bits in the transverse direction.

Claims

1. A method for determining coordinates of a circumscribed rectangular frame for enclosing each character based on image data of character images in a recognition area having been read by an image scanner, including:

a first step of determining specified image data of a predetermined data unit in a transverse direction parallel to lines including at least one bit of black pixel by scanning said read image data, and for determining a specified bit being one bit of black pixel in said specified image data;

a second step for judging whether a bit of black pixel is in either the bit located by one bit at the upper side of said specified bit of said specified image data which has been determined at said first step, or the bits located in said transverse direction by one bit of the bit located at the upper side by one bit of said specified bit of said specified image data;

a third step for determining the bit of black pixel as a newly specified bit if there is the bit of black pixel therein and for repeating said second step, otherwise the coordinate in the transverse direction of said specified bit of black pixel being determined as a top coordinate of said circumscribed rectangular frame;

a fourth step for judging whether a bit of black pixel is in either the bit located by one bit at the lower side of said specified bit of said specified image data which has been determined at said first step, or the bits located in said transverse direction by one bit of the bit located at the lower side by one bit of said specified bit of said specified image data;

a fifth step for determining the bit of black pixel as a newly specified bit if there is the bit of black pixel therein and for repeating said fourth step, otherwise the coordinate in the transverse direction of said specified bit of black pixel being determined as a bottom coordinate of said circumscribed rectangular frame;

a sixth step of calculating histogram data composed of either logical products or logical sums of respective bits in said transverse direction from plural image data located in a specified area between said determined top and bottom coordinates in a predetermined threshold range in said transverse direction from the center of said specified image data which is equal to or larger than the maximum of the transverse width of each circumscribed rectangular frame;

a seventh step of determining a coordinate in said transverse direction of a bit of black pixel located at the leftmost position in said calculated histogram data as a leftmost coordinate of said circumscribed rectangular frame; and

an eighth step of determining a coordinate in said transverse direction of a bit of black pixel located at the rightmost position in said calculated histogram data as a rightmost coordinate of said circumscribed rectangular frame.

2. The method as claimed in claim 1, further including a ninth step for judging said determined top and bottom coordinates and said determined leftmost and rightmost coordinates as coordinates of said circumscribed rectangular frame for enclosing each character when a difference between said determined top and bottom coordinates is equal to or smaller than a predetermined first threshold value and a difference between said determined leftmost and rightmost coordinates is equal to or smaller than a predetermined second threshold value after executing said eighth step.

3. The method as claimed in claim 2, further including:

a step of dividing said read image data into plural divided image data having predetermined data amount in said transverse direction so that said divided image data include image data overlapped with former divided image data adjacent thereto before executing said first step; and

a step of erasing bits of black pixel in said image data located within said determined circumscribed rectangular frame so as to set them at white pixels after executing said ninth step.

4. The method as claimed in claim 1, wherein histogram data is calculated only from logical products or logical sums of respective bits in the transverse direction.