System and method for automatically distinguishing between graphic information and text information of image data

• US 5,375,197 A
• Filed: 12/30/1991
• Issued: 12/20/1994
• Est. Priority Date: 12/31/1990
• Status: Expired due to Fees

• Current Assignee: Goldstar Jewellery LLC
• Original Assignee: Goldstar Jewellery LLC
• Inventors: Kang, Min S.
• Primary Examiner(s): Beausoliel, Jr., Robert W.
• Assistant Examiner(s): Elmore, Stephen C.
• Application Number:
  US07/816,040
• Time in Patent Office: 1,086 Days
• Field of Search: 395/147, 382/9, 382/11, 382/57, 382/61, 382/55
• US Class Current: 358/462
• CPC Class Codes: H04N 1/40062 Discrimination between different image types, e.g. two-tone, continuous tone

Abstract

A system for automatically distinguishing between graphic information and text information of image data, comprising a buffering circuit for storing image data from an A/D converter by an address from an address generator or outputting the stored image data to a data bus by an address from a microprocessor under a control of the microprocessor. Also, according to the present invention, there is provided a method for automatically distinguishing between graphic information and text information of image data, comprising the steps of converting an analog image signal from a scanner into digital image data and storing the digital image data by the line unit, reading the stored image data by the four line unit, subdividing the read image data into contiguous image data blocks and evaluating a block average, evaluating a background color illumination utilizing the evaluated block average, setting a distinguishing unit as windows each of which includes a predetermined number of contiguous blocks and distinguishing between a text portion and a graphic portion of the image data by the window unit, correcting an incorrectly distinguished window, performing an alignment of the edges of graphic/text portions, and correcting the delay of the real time and completing the distinction between graphic/text portions.

Claims

1. A method for automatically distinguishing between information and text information of image data, comprising the steps of:

(a) converting an analog image signal outputted from a scanner into digital image data and storing the digital image data for each line unit;

(b) reading the digital image data stored at said step (a) on the basis of four line units;

(c) subdividing the read image data into contiguous image data blocks with a desired mask and evaluating a block average;

(d) evaluating a background color illumination utilizing the block average evaluated at said step (c);

(e) setting a distinguishing unit as windows each of which includes a predetermined number of contiguous blocks and distinguishing between a text portion and a graphic portion of the image data by use of a window unit;

(f) correcting an incorrectly distinguished window after completing said step (e);

(g) performing an alignment of edges of graphic/text portions after the completion of the step (e); and

(h) correcting for a real time delay after the completion of the step (g) and then completing distinguishing between graphic/text portions;

wherein said step (e) of distinguishing between the text portion and the graphic portion of the image data includes the step of;

if the window includes at least one block with the background color illumination, determining the window as a window of the text portion of the image data;

if the window includes no block with the background color illumination, determining the window as a window of the graphic portion of the image data.

2. A method for automatically distinguishing between graphic information and text information of image data, as set forth in claim 1, wherein said block with the background color illumination is present in a block set of block average greater than a value obtained by multiplying the background color illumination by an error correcting value.

3. A method for automatically distinguishing between graphic information and text information of image data, as set forth in claim 2, wherein said error correcting value is a value below, but closely near one, which is obtained by experiment.

4. A method for automatically distinguishing between graphic information and text information of image data, comprising the steps of:

(a) converting an analog image signal outputted from a scanner into digital image data and storing the digital image data for each line unit;

(b) reading the digital image data stored at said step (a) on the basis of four line units;

(c) subdividing the read image data into contiguous image data blocks with a desired mask and evaluating a block average;

(d) evaluating a background color illumination utilizing the block average evaluated at said step (c);

(e) setting a distinguishing unit as windows each of which includes a predetermined number of contiguous blocks and distinguishing between a text portion and a graphic portion of the image data by use of a window unit;

(f) correcting an incorrectly distinguished window after completing said step (e);

(g) performing an alignment of edges of graphic/text portions after the completion of the step (e); and

(h) correcting for a real time delay after the completion of the step (g) and then completing distinguishing. between graphic/text portions;

wherein said step (f) of correcting the incorrectly distinguished window includes the step of;

in vertical and horizontal correction cases where both edge windows of vertical and horizontal contiguous five windows are those of a same portion and at least one of inner windows is one of a portion different from a portion of edge windows, correcting the window of the different portion as a window of the same portion as that of the edge windows.

5. A method for automatically distinguishing between graphic information and text information of image data, comprising the steps of:

(a) converting an analog image signal outputted from a scanner into digital image data and storing the digital image data for each line unit;

.

(b) reading the digital image data stored at said step (a) on the basis of four line units;

(c) subdividing the read image data into contiguous image data blocks with a desired mask and evaluating a block average;

(d) evaluating a background color illumination utilizing the block average evaluated at said step (c);

(e) setting a distinguishing unit as windows each of which includes a predetermined number of contiguous blocks and distinguishing between a text portion and a graphic portion of the image data by use of a window unit;

(f) correcting an incorrectly distinguished window after completing said step (e);

(g) performing an alignment of edges of graphic/text portions after the completion of the step (e); and

(h) correcting for a real time delay after the completion of the step (g) and then completing distinguishing between graphic/text portions;

wherein said step (g) of performing the alignment edges of graphic/text portions includes the step of;

thinning a window representative of a varied information portion which is changed from the graphic portion to the text portion or from the text portion to the graphic portion, to align an edge of the window.

6. A method for automatically distinguishing between graphic information and text information of image data, as set forth in claim 5, wherein said thinning step includes the step of:

discriminating whether each block of the window representative of the varied information portion has the illumination higher than a threshold, discrimination is started with left end blocks and distinguishing the blocks having the illumination higher than the threshold as the text portion.