| Input: Document under scrutiny, , watermarked keys; , |
| , fixed set-size, = 2. |
| Output: Validity decision, output image logo that |
| identifies the owner. |
| (1) Divide into -word sets according to the . |
| (2) for = 1 to |
| get -th word-set from . |
| apply to add the Unicode values of all |
| characters in each word of set- to produce |
| and |
| convert and into binary values. |
| end for |
| (3) Embed into by performing a logical |
| NAND-operation, to produce an extracted key, . |
| (2) Compare with the to test for equivalence, . |
| (3) if ( == 1) then |
| while bits left in -result, in do |
| Find the bit-pattern of the first-words of each set, |
| using knowledge from the valid-document, , of the |
| binary values of the second words in each word-set |
| together with generation of possible inputs of the |
| logical-XOR operation (from encoder)—as in the |
| -results found in -to identify the unknown bits of |
| the first-word within each word-set. |
| Convert the derived first-word bit-results () and the |
| first-word bit-results of () into decimal. |
| Perform to obtain the Unicode-decimal values. |
| Determine the embedded characters in each of the |
| first-words of from the Unicode-values. |
| end while |
| end if |
| (4) Convert extracted characters to produce an image , |
| using the Text-To-Image convertor. |
| (5) Output logo or invalid-decision. |
