초록 ▼

A document is analyzed to determine whether it is a banknote or the like, by reference to image data corresponding to the document. Two analysis techniques may be used, one based on detection of a visible pattern characteristic of a security document, the other based on detection of a steganographic

A document is analyzed to determine whether it is a banknote or the like, by reference to image data corresponding to the document. Two analysis techniques may be used, one based on detection of a visible pattern characteristic of a security document, the other based on detection of a steganographic digital watermark characteristic of a security document. If either characteristic is found, the image is flagged, and appropriate anti-counterfeiting steps may be taken. Detection of the visible pattern can be performed using a series of successively more rigorous tests. If the image fails a test, successive testing steps can be skipped, speeding the process. Hough transform-based pattern recognition techniques are used in some embodiments. Provision of both a visible pattern detector and a watermark detector in a single apparatus enhances detection reliability, while permitting various implementation efficiencies.

대표청구항 ▼

1. A method of processing image data to screen for security document images comprising, in the order stated: (a) performing a first analysis on the image data; (b) if the first analysis indicates the image data does not correspond to a security document, skipping steps (c)-(e) (c) performing a

1. A method of processing image data to screen for security document images comprising, in the order stated: (a) performing a first analysis on the image data; (b) if the first analysis indicates the image data does not correspond to a security document, skipping steps (c)-(e) (c) performing a second analysis on the image data; (d) if the second image analysis indicates the image data does not correspond to a security document, skipping step (e); and (e) flagging the image data as corresponding to a security document; wherein at least one of the analyses employs a Hough-based transform to detect a two-dimensional, rotationally invariant feature. 2. The method of claim 1 wherein the image data is acquired by visible light scanning. 3. The method of claim 1 in which one of the analyses employs a Hough-based transform to detect a two-dimensional, rotationally invariant feature at a first rotational state, and another of the analyses employs a Hough-based transform to detect a two-dimensional, rotationally invariant feature at a second, different, rotational state. 4. A method of processing image data to screen for security document images comprising, in the order stated: (a) performing a first analysis on the image data; (b) if the first analysis indicates the image data does not correspond to a security document, skipping steps (c)-(e) (c) performing a second analysis on the image data; (d) if the second image analysis indicates the image data does not correspond to a security document, skipping step (e); and (e) flagging the image data as corresponding to a security document; wherein the first analysis is based on a rotationally invariant feature. 5. The method of claim 4 wherein the image data is acquired by visible light scanning. 6. A method of processing image data to screen for security document images comprising, in the order stated: (a) performing a first analysis on the image data; (b) if the first analysis indicates the image data does not correspond to a security document, skipping steps (c)-(e) (c) performing a second analysis on the image data; (d) if the second image analysis indicates the image data does not correspond to a security document, skipping step (e); and (e) flagging the image data as corresponding to a security document; wherein the first analysis is based on a rotationally symmetric feature. 7. The method of claim 6 wherein the image data is acquired by visible light scanning. 8. A method of processing image data to screen for security document images comprising, in the order stated: (a) performing a first analysis on the image data; (b) if the first analysis indicates the image data does not correspond to a security document, skipping steps (c)-(e) (c) performing a second analysis on the image data; (d) if the second image analysis indicates the image data does not correspond to a security document, skipping step (e); and (e) flagging the image data as corresponding to a security document; wherein one of said analyses steps through plural rotation states of the image data.