Method of searching and extracting text information from drawings

• US 5,995,659 A
• Filed: 09/09/1997
• Issued: 11/30/1999
• Est. Priority Date: 09/09/1997
• Status: Expired due to Term

• Current Assignee: Siemens Corp. (Siemens AG)
• Original Assignee: Siemens Corporate Research Incorporated (Siemens AG)
• Inventors: Hsu, Liang H., Chakraborty, Amit, Liu, Peiya
• Primary Examiner(s): Tran, Phuoc
• Application Number:
  US08/926,217
• Time in Patent Office: 812 Days
• Field of Search: 382/172, 382/176, 382/177, 382/190, 382/202, 382/205, 382/262, 382/281, 382/171, 382/193, 382/197, 358/462, 358/464, 358/453, 707/501, 707/502, 707/513, 707/530, 707/531, 707/504
• US Class Current: 382/176
• CPC Class Codes: G06V 30/413 Classification of content, e.g. text, photographs or tables

Abstract

A method for recognizing text in graphical drawings includes creating a binarized representation of a drawing to form an electronic image of pixels. The image is discriminated between text regions and lines in the image by grouping pixels into blocks and comparing blocks with a predetermined format to identify text regions. The lines that remain in the text regions are removed to create text only regions. The text is recognized in the text only regions.

Claims

1. A method for recognizing text in graphical drawings comprising the steps of:

creating a binarized representation of a drawing to form an electronic image of pixels;

discriminating text regions from lines in the image by determining a distance between each black pixel and nearest neighboring black pixels, grouping pixels into pixel blocks having a same label based on a relationship between the determined distance and a predetermined distance, and comparing the blocks with a predetermined format to identify text regions;

removing the lines from the image using a Hough Transform to create text only regions; and

recognizing text in the text only regions.

2. The method as recited in claim 1 further comprising the step of optically scanning the drawing.

3. The method as recited in claim 1 wherein the step of creating a binarized representation includes the steps of comparing pixel values to a grey-scale threshold value and assigning the pixels to be either white or black.

4. The method as recited in claim 1 wherein the step of discriminating text regions includes the steps of:

comparing each distance to the predetermined distance; and

assigning labels to pixels within the predetermined distance to form the pixel blocks having the same label.

5. The method as recited in claim 1 wherein the step of removing the lines includes the steps of:

subdividing the drawing space into a grid defining grid spaces;

counting black pixels associated with text and graphics lines in each grid space;

comparing a count of black pixels to a predetermined count for determine if the grid space contains a line; and

removing lines by changing black pixels to white pixels.

6. The method as recited in claim 1 wherein the step of recognizing text in the text only regions includes using optical character recognition software.

7. The method as recited in claim 1 further comprising the step of cleaning up the image by median filtering.

8. The method as recited in claim 1 further comprising the step of creating anchorable information units from the text only regions for hypertext creation.

9. A method for recognizing text in graphical drawings comprising the steps of:

creating a binarized representation of a drawing to form an electronic image of black pixels and white pixels;

determining a distance between each black pixel and nearest neighhboring black pixels;

comparing each distance to a predetermined distance;

assigning labels to pixels within the predetermined distance to form pixel blocks having the same label; and

comparing pixel blocks to predetermined text formats for identifying text regions;

removing the lines from the image using a Hough Transform to create text only regions; and

recognizing text in the text only regions.

10. The method as recited in claim 9 further comprising the step of optically scanning the drawing.

11. The method as recited in claim 9 wherein the step of creating a binarized representation includes the steps of comparing pixel values to a grey-scale threshold value and assigning the pixels to be either white or black.

12. The method as recited in claim 9 wherein the step of removing the lines incudes the steps of:

subdividing the drawing space into a grid defining grid spaces;

counting black pixels associated with text and graphics lines in each grid space;

comparing a count of black pixels to a predetermined count for determine if the grid space contains a line; and

removing lines by changing black pixels to white pixels.

13. The method as recited in claim 9 wherein the step of recognizing text in the text only regions includes using optical character recognition software.

14. The method as recited in claim 9 further comprising the step of cleaning up the image by median filtering.

15. The method as recited in claim 9 further comprising the step of creating anchorable information units from the text only regions for hypertext creation.

16. A method for recognizing text in graphical drawings comprising the steps of:

creating a binarized representation of a drawing to form an electronic image of black pixels and white pixels;

determining a distance between each black pixel and nearest neighhboring black pixels;

comparing each distance to a predetermined distance;

assigning labels to pixels within the predetermined distance to form pixel blocks having the same label; and

comparing pixel blocks to predetermined text formats for identifying text regions;

subdividing the drawing space into a grid defining grid spaces;

counting black pixels associated with text and graphics lines in each grid space;

comparing a count of black pixels to a predetermined count for determine if the grid space contains a line;

removing lines by changing black pixels to white pixels to create text only regions;

recognizing text in the text only regions; and

converting the text to hypertext for creating an anchorable information unit.

17. The method as recited in claim 16 further comprising the step of optically scanning the drawing.

18. The method as recited in claim 16 wherein the step of creating a binarized representation includes the steps of comparing pixel values to a grey-scale threshold value and assigning the pixels to be either white or black.

19. The method as recited in claim 16 wherein the step of recognizing text in the text only regions includes using optical character recognition software.

20. The method as recited in claim 16 further comprising the step of cleaning up the image by median filtering.