US 7,644,281 B2 · Character and vector graphics watermark for structured electronic documents security

Abstract

The present invention is a method and apparatus for watermarking text or vector graphics documents. It is based on character-wise or vector graphics element-wise grayscale or color modulation. At high resolution, halftone or dither modulation can also be used in addition or in place of grayscale/color modulation, for printed representation of electronic document. For the detection, the document is acquired through an acquisition device, document segmentation is performed, characters/elements are segmented, the watermark signal is estimated and the information is decoded. Although the proposed scheme mostly addresses the watermarking of hard-copy documents, it can easily be integrated into electronic document editing and acquisition tools, and the watermark is attached to the document electronic version. The invention is applicable either using expensive high-resolution printing and acquisition devices, either using common cheap low-resolution devices, depending on the application needs. The proposed scheme is suitable for example to the protection of security documents, contracts, technical and commercial documentation; it can use any physical support like paper, cellulose, or plastic; it can be used for copy protection, authentication, or tamper proofing; finally it can also be applied to other non security-related applications, such as document tracking, as well as document embedded annotation and watermarked-assisted automatic processing.

Claims23

Dependent claims

A method for embedding a digital watermark w into textual and/or vector graphics documents of electronic and/or hard-copy form, based on the modulation of luminance or grayscale values, of color values, and/or of halftone patterns of characters and/or of vector elements, the method comprising the steps of

(a) encoding a digital message m, resulting into a codeword c;
spatially allocating the codeword c, resulting into a watermark signal w suitable for text and/or vector graphics modulation,
(b) parsing the original input document in electronic format x in order to isolate all its components,
(c) selecting text and/or vector graphics components xtext which will carry the watermark w and parsing components xtext in order to select single entities or elements to use to encode individual w symbols;
and keeping the remaining non selected components xnon-text for later document recomposition,
(d) modulating the grayscale attributes, the color attributes and/or the halftone pattern attributes of the selected elements in the selected text and/or vector graphics components xtext,
(e) recomposing the document from the marked version of selected components xtext and the remaining components xnon-text, resulting into a watermarked document y, which can be stored, distributed, and/or used in electronic form as electronic document {hacek over (y)}a or in printed form as hard copy document {hacek over (y)}b.

4 more elements

A method for extracting a digital watermark ŵ from textual and/or vector graphics documents of electronic and/or hard-copy form based on the modulation of luminance or grayscale values, of color values, or of halftone patterns of characters or of vector elements, wherein the watermarked and potentially distorted or attacked document {hacek over (y)}′ is segmented, processed and the watermark extracted as ŵ, the codeword ĉ estimated, and the message {circumflex over (m)} decoded, the method comprising the steps of

(a) applying document segmentation to the watermarked document {hacek over (y)}′, including the selection of documents text and/or vector graphics components ytext,
(b) extracting the potentially modulated elements from the text and/or vector graphics components, including performing character and/or elements segmentation for text and/or vector graphics documents,
(c) estimating the watermark signal ŵ from the modulated grayscale, color and/or halftone pattern attributes,
(d) generating a pilot signal;
estimating and compensating the extracted watermark ŵ state and desynchronization, based on the pilot signal, resulting into the estimated codeword ĉ;
decoding the codeword ĉ resulting into the final message {circumflex over (m)}.

3 more elements

Dependencies are inferred from the claim text.

Character and vector graphics watermark for structured electronic documents security

Life of this patent

Character and vector graphics watermark for structured electronic documents security

Life of this patent

Abstract

Claims23