Methods and systems for watermarking documents
Described are methods and system that digitally watermark documents using an encoding scheme that embeds unique identifiers into the files by perturbing the relative heights of words to create patterns that are imperceptible to the human eye but decodable to recover the identifiers. The unique identifiers can be precoded before they are embedded in the documents.
1. A method for encoding a user identifier in words of a document, the method comprising:
capturing original images of the words of the document, each of the original images having respective renderable dimensions encompassing a word area;
changing the renderable dimensions of each of the word areas to create a scaled area corresponding to each of the word areas, the changed renderable dimensions of the scaled areas encoding the user identifier into the scaled areas; and
replacing or covering the original images of the words with the corresponding scaled areas.
2. The method of claim 1 , wherein capturing the original image of each of the words comprises delineating a bounding box around the word, the bounding box encompassing an original area that includes the word and background content.
3. The method of claim 1 , wherein the background content includes a margin around the word to cover the word when the original area is scaled down.
4. The method of claim 2 , wherein the extracting comprises recognizing the words.
5. The method of claim 1 , wherein encoding the user identifier in the words of the document comprises associating the document with a document identifier, pre-encoding the user identifier with the document identifier into a bitstream, and encoding the bitstream into the scaled areas of the document.
6. The method of claim 5 , wherein the pre-encoding maps each character of the user identifier to a variable-length bit sequence.
7. The method of claim 5 , further comprising extracting the bitstream from the document.
8. The method of claim 7 , wherein the document is a leaked version of an original document identified by the document identifier, the method further comprising decoding the document identifier from the bitstream, comparing the leaked version with the original document, and determining whether the leaked version is a portion of the original document.
9. The method of claim 5 , further comprising encoding instances of the bitstream into the scaled areas of the document.
10. The method of claim 9 , wherein scaled areas are of maximum and minimum heights, the method further comprising separating the instances of the bitstream with words of an average of the maximum and minimum heights that serve as an end marker.
11. The method of claim 9 , wherein the document includes pages, and wherein every page includes at least a portion of the bitstream, an end marker, or both.
12. A method of identifying a user associated with a document, the method comprising:
detecting lines of words in the document, the lines and the words having respective line dimensions and word dimensions, including word heights, word widths, line heights, and line widths;
normalizing the word heights relative to the line heights and the word widths relative to the line widths;
scaling the word dimensions responsive to the normalized word heights and normalized word widths; and
decoding a user identifier from the scaled word dimensions.
13. The method of claim 12 , further comprising scanning at least a portion of the document to extract the lines of words.
14. The method of claim 12 , wherein the document is associated with a document identifier, the method further comprising decoding the decoded user identifier using the document identifier.
15. The method of claim 14 , wherein decoding the decoded user identifier using the document identifier comprises extracting a bitstream from the dimensions and extracting the user identifier from the bitstream with the document identifier.
16. A method for encoding at least a portion of a user identifier in a line of words of a line height, the method comprising:
capturing original images of the words, each of the original images having a word height;
changing the heights of the original images to create scaled word areas collectively encoding the user identifier; and
replacing or covering the original images of the words with the scaled word areas.
17. The method of claim 16 , wherein each of the original images includes a margin for the imaged word and the scaled areas of at least a subset of the scaled areas are smaller than the original images.
18. The method of claim 17 , wherein the margins cover the words in the original images.
19. The method of claim 16 , wherein the line of words is of a line width and each of the original images have a word width, the method further comprising:
normalizing the word heights, wherein each of the normalized word heights is the ratio between the word height and the line height; and
changing the normalized heights of the original images to create the scaled word areas.