Watermarking image block division method and device for western language watermarking processing
The application provides a method for partitioning a watermark image with western language characters, comprising: partitioning a western language characters image along rows and columns to form a plurality of character image blocks; identifying valid character image blocks from the formed character image blocks; counting sizes of the valid character image blocks to determine if the image corresponds to a document with a large font size or a document with a small font size; dividing words in the image into a plurality of groups, wherein each divided group in the document with large font size has different numbers of words from that with small font size; and dividing equally the divided word groups into multiple portions corresponding to watermark image blocks. The application further provides a device for partitioning a watermark image with western language characters. The operability of watermark embedding process can be ensured through the above technical solution.
1. A method for partitioning a watermark image with western language characters, comprising:
partitioning a western language characters image along rows and columns to form a plurality of character image blocks;
identifying valid character image blocks from the formed character image blocks;
counting sizes of the valid character image blocks to determine if the image corresponds to a document with a large font size or a small font size;
dividing words in the image into a plurality of groups, wherein each divided group in the document with large font size has different numbers of words from that with small font size; and
dividing equally the divided word groups into multiple portions corresponding to watermark image blocks.
2. The method according to claim 1 , wherein the step of identifying comprises:
differentiating the formed character image blocks into punctuation image blocks and valid characters image blocks.
3. The method according to claim 2 , wherein the step of differentiating comprises:
judging if following conditions are satisfied:
condition 1: w>N t1 ×H;
condition 2: both of a top edge and a bottom edge of U fall into one side of m;
condition 3: the top edge and bottom edge of U respectively fall into two sides of m, and w<N t2 ×H;
where, U is a circumscribed rectangular box of a character image block in a collection Ω for character image blocks; H is a row height of the row where U is located; m is a center line of the row; h and w respectively are a height and a width of U; and both of N t1 and N t2 are preset coefficients,
determining that each of the formed character image blocks, which corresponds to U, is a punctuation image block if any one of the three conditions is satisfied; otherwise
determining that each of the character image blocks corresponding to U is a valid character image block.
4. The method according to claim 3 , wherein N t1 =4 and N t2 =0.35.
5. The method according to claim 3 , wherein the step of counting comprises a step of calculating Hs by rule of
Hs
=
h
1
+
h
2
+
…
+
h
m
m
,
where, h 1 , h 2 , . . . , h m respectively are heights of the valid character image blocks 1, 2, . . . , m in the row;
determining that the row has large font size if Hs≧Th size ;
determining that the row has small font size if Hs<Th size , where Th size is a preset threshold;
counting the number N large of the row with large font size and the number N small of the row with small font size;
determining that the image corresponds to the document with large font size if N large ≧N small ; and
determining that the image corresponds to the document with small font size if N large <N small .
6. The method according to claim 5 , wherein Th size =88.
7. The method according to claim 1 , wherein the step of dividing equally the divided word groups comprises:
dividing equally the word groups into fixed number of portions according to a valid length projected from a column of words in the word groups; and
forming a circumscribed rectangular box of each of the portions by its width and maximum height of characters included in each of the portions, wherein the circumscribed rectangular box corresponds to the watermark image block.
8. The method according to claim 1 , wherein, before the step of partitioning, the method further comprises:
obtaining the western language characters image; and
de-noising the obtained image to form a binary image.
9. A device for partitioning a watermark image with western language characters, comprising:
a partitioning module configured to partition a western language characters image along rows and columns to form a plurality of character image blocks;
an identifying module configured to identify valid character image blocks from the formed character image blocks;
a statistic module configured to count sizes of the valid character image blocks to determine if the western language characters image has a large font size or a small font size;
a dividing module configured to divide words in the image into a plurality of groups, wherein each divided group for the document with large font size has different numbers of words from that with small font size; and
an equally dividing module configured to equally divide the divided word groups into multiple portions corresponding to watermark image blocks.
10. The device according to claim 9 , wherein the identifying module comprises a judging module configured to determine if following conditions are satisfied:
condition 1: w>N t1 ×H;
condition 2: both of a top edge and a bottom edge of U fall into one side of m;
condition 3: the top edge and bottom edge of U respectively fall into two sides of m, and w<N t2 ×H;
where, U is a circumscribed rectangular box of a character image block in a collection Ω for character image blocks; H is a row height of the row where U is located; m is a center line of the row; h and w respectively are a height and a width of U; and both of N t1 and N t2 are preset coefficients,
a determining module configured to determine that each of the formed character image blocks, which corresponds to U, is a punctuation image block if any one of the three conditions is satisfied; otherwise each of the character image blocks corresponding to U is a valid character image block.
11. The device according to claim 10 , wherein the statistic module comprises a calculating module configured to calculate Hs by rule of
Hs
=
h
1
+
h
2
+
…
+
h
m
m
,
where, h 1 , h 2 , . . . , h m respectively are heights of the valid character image blocks 1, 2, . . . , m in the row;
a row determining module configured to determine that the row has large font size if Hs≧Th size , otherwise the row has small font size, where the Th size is a preset threshold;
a counting module configured to count the number N large of the row with large font size and the number N small of the row with small font size; and
a document determining module configured to determine that the western language character image corresponds to the document with large font size if N large ≧N small , and to determine the western language character image corresponds to the document with small font size if N large <N small .