Methods Combining Multiple Frames of Image Data

• US 20080181450A1
• Filed: 03/17/2008
• Published: 07/31/2008
• Est. Priority Date: 05/02/2000
• Status: Active Grant

• Current Assignee: Digimarc Corporation
• Original Assignee: Digimarc Corporation
• Inventors: Bradley, Brett Alan, Rhoads, Geoffrey B., Carr, J. Scott
• Granted Patent: US 8,126,272 B2
• Time in Patent Office: Days
• US Class Current: 382/100
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/1448 based on markings or identifiers characterising the document or the area
• CPC Class Codes: G07D 7/0047 using checkcodes, e.g. coded numbers derived from serial number and denomination
• CPC Class Codes: G07D 7/005 Testing security markings invisible to the naked eye, e.g. verifying thickened lines or unobtrusive markings or alterations

Abstract

In one embodiment a document authentication station, for use with passports or the like, includes a 2D image sensor (e.g., CCD- or CMOS-based video camera), and a computer device. The image sensor produces produce image data corresponding to a presented document. From this image data, the computer extracts two or more identification data. One is a digital watermark. The other can be a bar code, data glyphs, OCR data, etc. The processor then proceeds to check that the two identification data correspond in an expected fashion. If not, the document is flagged as suspect or fake. Reliability of detection can be enhanced by processing plural frames of data from the image sensor before issuing a result.

Claims

1. A method comprising the acts:

at a first time, capturing a first set of image data that depicts a text-bearing physical subject;

at a second, subsequent, time, capturing a second set of image data that depicts said subject, the second set of image data being different than the first set of image data;

identifying a portion of one of said sets of image data that is unreliable for recognizing text from the text-bearing object; and

processing said sets of image data to recognize text therefrom;

wherein said processing takes into consideration that a portion of one of the sets of image data is unreliable.

2. The method of claim 1 wherein said identifying comprises identifying a portion that suffers from glare.

3. The method of claim 1 wherein said identifying comprises identifying a portion that is not in focus.

4. The method of claim 1 wherein said processing comprises generating a new set of image data from said first and second sets of image data, taking into consideration that a portion of one of the sets of image data is unreliable.

5. The method of claim 4 that includes generating a composite set of image data, said composite set of image data omitting data from the portion identified as unreliable.

6. The method of claim 1 that further includes sensing digital watermark data from said physical subject, and using the digital watermark data in conjunction with the recognized text to make a decision concerning said subject, wherein both the capturing and sensing make use of a digital video camera that captures successive frames of image data, and the subject comprises a photo identification document.

7. The method of claim 6 that includes determining if the digital watermark data and the recognized text correspond in an expected fashion and, if not, flagging the photo identification document as suspect.

8. A method comprising the acts:

at a first time, capturing a first set of image data that depicts a text-bearing physical subject;

at a second, subsequent, time, capturing a second set of image data that depicts said subject, the second set of image data being different than the first set of image data;

identifying a region in the first set that is relatively more reliable for recognizing text than another region thereof;

identifying a region in the second set that is relatively more reliable for recognizing text than another region thereof;

combining image data from the reliable region in the first frame with image data from the reliable region in the second frame; and

performing an optical character recognition operation to recognize text from said combined image data.

9. The method of claim 8 that further includes sensing digital watermark data from said physical subject, and using the digital watermark data in conjunction with the recognized text to make a decision concerning said subject.

10. The method of claim 9 that includes determining if the digital watermark data and the recognized text correspond in an expected fashion and, if not, flagging the subject as suspect.

11. The method of claim 9 that includes consulting a database to obtain additional data corresponding to at least a portion of said watermark data, and comparing said additional data with data corresponding to the recognized text.