Image analysis – Applications
Reexamination Certificate
2001-07-20
2004-11-23
Mehta, Bhavesh M. (Department: 2621)
Image analysis
Applications
C382S232000, C380S054000, C713S176000
Reexamination Certificate
active
06823076
ABSTRACT:
FIELD OF THE INVENTION
The invention relates generally to the field of digital image processing, and in particular to the field of embedding digital data in an original image in such a manner that the embedded data can be completely removed at a later time to allow lossless recovery of the original image.
BACKGROUND OF THE INVENTION
In some digital imaging systems, it is desirable to convey ancillary information along with the actual data that comprises an original image. This ancillary information might be directly derived from the image, or it might represent additional information that is related to, but not directly derived from, the image itself. In either case, this ancillary information is called image “meta-data” throughout this text.
An example of meta-data that is directly derived from the original image data is a hash value. A hash value is a very compact representation of the much larger original image data set, and it is generated using a hashing function. An example of a useful hashing function is found in U.S. Department of Commerce Technology Administration National Institute of Standards and Technology, FIPS PUB 180, “Secure Hash Standard,” May 11, 1993, pp. 1-20. This hashing function will produce a hash value of length 162 bits, irrespective of the number of pixel values that are input to the hashing function. Other hashing functions may produce hash values of lengths other than 162 bits. While a hash value is not entirely unique to a given image (i.e., the process is a many-to-one mapping), it can be chosen so as to represent the image with extremely high probability. A hash value can be appended to the original image and then used at a later time to verify that the image has not been modified in any way since the hash value was generated. To prevent tampering with the hash value by unauthorized individuals, it is necessary to encrypt the hash value, thus creating a secure digital signature for the original image data. However, the process of appending the encrypted hash value inhibits the use of standard image formats such as TIFF to convey the combined image and signature information. Furthermore, a signature that is merely appended to an image can be easily removed by deleting the portion of the file containing the signature.
Examples of image meta-data that are not directly derived from, but are related to, the original image include the date/time or geographical location of the point of capture, a unique ID associated with the camera and/or photographer, camera settings used during the capture process, etc. It is possible to merely append such meta-data to the image data, but as with the digital signatures this process inhibits the use of standard image file formats. Some file formats provide segments for user-defined information, but such data is unsecured from tampering and could be easily removed by unauthorized individuals. In present techniques, these issues are resolved by embedding the meta-data within the image itself. Standard image file formats such as TIFF can then be used to convey both image data and meta-data, and proper design of the embedding process allows the image containing the embedded data to be directly viewed with only a minimal loss in quality. The embedding process also provides some level of security in that the meta-data can only be recovered with some knowledge of the embedding process (such as a “key” value). However, a disadvantage of the embedding process is that the original image data is typically corrupted to the extent that exact recovery of the original values is impossible. As a result, many current data embedding techniques could not be used for the purpose of image verification where the encrypted hash values (i.e., image signatures) are embedded in the image itself (since the proper hash value could never be re-created from the corrupted image values).
Recent techniques have been proposed that address this issue (see commonly assigned U.S. patent application Ser. No. 09/074,282 entitled “Lossless Recovery of an Original Image Containing Embedded Data”, which was filed May 7, 1998 in the names of Honsinger, C., Jones, P., Rabbani, M., and Stoffel, J.; an article by Fridrich, J., Goljan M., and Du, R.,“Invertible Authentication”,
Proc. SPIE Security and Watermarking of Multimedia Contents
, January, 2001; and an article by Goljan, M., Fridrich, J., and Du, R., “Distortion-free Data Embedding for Images”,
Proc
. 4
th Information Hiding Workshop
, April, 2001). Each describes embedding methods by which the original image values can be recovered exactly after extraction of the embedded data. For example, in one method described in each of these sources, the embedded data may be combined with the original image using a reversible transform, e.g., a modulo-N addition, to form a digital image containing the embedded data. For general imaging systems, these data embedding methods will suffice to re-create the original image.
If a scannerless range imaging system is considered, a new data embedding method can be used that provides advantages over all of the current data embedding methods described for general imaging systems. U.S. Pat. No. 4,953,616 describes a scannerless range imaging system (further described in the Sandia Lab News, vol. 46, No. 19, Sep. 16, 1994) using either an amplitude-modulated high-power laser diode or an array of amplitude-modulated light emitting diodes (LEDs) to completely illuminate a target scene. An improved scannerless range imaging system that is capable of yielding color intensity images in addition to the 3D range images is described in commonly-assigned, copending U.S. patent application Ser. No. 09/572,522, filed May 17, 2000 and entitled “Method and Apparatus for a Color Scannerless Range Imaging System”. (As used herein, a scannerless range imaging (SRI) system will sometimes be referred to as an “SRI camera”.) In the formation of a three-dimensional image (which will herein refer to the combination of the intensity image and the range image), the SRI camera generates an “image bundle”, which includes both the intensity image and a collection of phase images which are used to construct the range image.
It would be desirable to have an embedding technique that would incorporate the above-mentioned advantages in a three-dimensional image of the type captured by an SRI camera. However, there are several problems that need to be addressed. In particular, it would be desirable to embed any meta-data in the three-dimensional image formed by an SRI camera such that the range data can be exactly reconstructed from the altered image without having to extract the embedded data. Moreover, it would be beneficial if the data does not have to be embedded in the intensity or range images (or the three-dimensional image), so the original image values of the intensity and range images are not altered.
SUMMARY OF THE INVENTION
The present technique is an improvement based on the technique of data embedding that is described in commonly assigned U.S. Pat. No. 5,859,920, entitled, “Method for Embedding Digital Information in an Image” by Daly et al., and as modified in commonly assigned U.S. Pat. No. 6,044,156, entitled, “Method for Generating an Improved Carrier for use in an Image Data Embedding Application” by Honsinger et al., both of which are incorporated herein by reference. More specifically, the present invention extends these data embedding techniques for use in a scannerless range imaging system. However, in a departure from the prior art, the embedded data is combined with the phase image(s) rather than the original pictorial image. As will be further seen, this leads to unexpected benefits in the recovery of the phase images themselves.
The present invention is directed to overcoming one or more of the problems set forth above. Briefly summarized, according to one aspect of the present invention, the invention resides in a method for embedding data into the output of a scannerless range imaging system of the type that includes an illumination system for controllably illuminating a scene
Cahill Nathan D.
Ray Lawrence A.
Bayat Ali
Eastman Kodak Company
Mehta Bhavesh M.
Woods David M.
LandOfFree
Method for embedding digital information in a three... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method for embedding digital information in a three..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for embedding digital information in a three... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3355954