GB2374478A - Restoring compromised image data - Google Patents

Restoring compromised image data Download PDF

Info

Publication number
GB2374478A
GB2374478A GB0108427A GB0108427A GB2374478A GB 2374478 A GB2374478 A GB 2374478A GB 0108427 A GB0108427 A GB 0108427A GB 0108427 A GB0108427 A GB 0108427A GB 2374478 A GB2374478 A GB 2374478A
Authority
GB
United Kingdom
Prior art keywords
video images
images
image
indicative
video
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Withdrawn
Application number
GB0108427A
Other versions
GB0108427D0 (en
Inventor
Jia Hong Yin
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Central Research Laboratories Ltd
Original Assignee
Central Research Laboratories Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Central Research Laboratories Ltd filed Critical Central Research Laboratories Ltd
Priority to GB0108427A priority Critical patent/GB2374478A/en
Publication of GB0108427D0 publication Critical patent/GB0108427D0/en
Publication of GB2374478A publication Critical patent/GB2374478A/en
Withdrawn legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T1/00—General purpose image data processing
    • G06T1/0021—Image watermarking
    • G06T1/005—Robust watermarking, e.g. average attack or collusion attack resistant
    • G06T1/0064—Geometric transfor invariant watermarking, e.g. affine transform invariant
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T1/00—General purpose image data processing
    • G06T1/0021—Image watermarking
    • G06T1/0028—Adaptive watermarking, e.g. Human Visual System [HVS]-based watermarking
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
    • H04N1/00002—Diagnosis, testing or measuring; Detecting, analysing or monitoring not otherwise provided for
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
    • H04N1/00002—Diagnosis, testing or measuring; Detecting, analysing or monitoring not otherwise provided for
    • H04N1/00005—Diagnosis, testing or measuring; Detecting, analysing or monitoring not otherwise provided for relating to image data
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
    • H04N1/00002—Diagnosis, testing or measuring; Detecting, analysing or monitoring not otherwise provided for
    • H04N1/00026—Methods therefor
    • H04N1/00037—Detecting, i.e. determining the occurrence of a predetermined state
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
    • H04N19/136—Incoming video signal characteristics or properties
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/46—Embedding additional information in the video signal during the compression process
    • H04N19/467—Embedding additional information in the video signal during the compression process characterised by the embedded information being invisible, e.g. watermarking
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
    • H04N19/59—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving spatial sub-sampling or interpolation, e.g. alteration of picture size or resolution

Abstract

Apparatus for restoring compromised image data, including recognizing watermark coding impressed upon video images received by way of a communication channel and subject to alteration prior to and/or during transfer through said channel, includes means for detecting characteristics of said images indicative of at least one feature of said alteration, and means for utilizing said detected characteristics to operate upon the video images received by way of said channel in a sense tending to restore said video images toward their unaltered form prior to their application to said recognition means. Preferably the means for detecting characteristics is responsive to linear boundaries between said video images and their surroundings, and the characteristics are indicative of shrinkage of said video images in both vertical and horizontal directions thereof.

Description

<Desc/Clms Page number 1>
RESTORING COMPROMISED IMAGE DATA This invention is concerned with restoring image data compromised by editing prior to and/or during transmission through a communications channel such as may be used, for example, for television broadcasting or for transmission over the Internet.
In particular, the invention is concerned with improving the ability of identification codes, or so-called "watermarks", inserted into visual image data for purposes such as indicating copyright ownership and/or inhibiting unauthorized copying or other treatment of the image data, to be detected accurately despite being compromised by editing imposed upon the visual image data.
Such editing might take a number of forms, but a typical and commonly used example is that of compressing the image horizontally and/or vertically so as to leave a blank area in which other data, such as sub-titles or other video images, may be inserted so as to be superimposed upon the transmitted data without actually overlapping the pictorial content of the image. In these circumstances, difficulty can arise in that a monitoring circuit, set up to detect watermark coding incorporated into the original video image, may not recognize the watermark code, and may therefore not respond in the correct manner to the reception of the edited image.
This invention seeks to address the above-described difficulty.
According to the invention there is provided recognition means for recognizing watermark coding impressed upon video images received by way of a communication channel and subject to alteration prior to and/or during transfer
<Desc/Clms Page number 2>
through said channel, including means for detecting characteristics of said images indicative of at least one feature of said alteration, and means for utilizing said detected characteristics to operate upon the video images received by way of said channel in a sense tending to restore said video images toward their unaltered form prior to their application to said recognition means.
Preferably, said means for detecting characteristics is responsive to linear boundaries between said video images and their surroundings. This permits frequently used editorial processes such as image shrinkage to be readily recognized. Frequently, said linear boundaries run horizontally and are indicative of shrinkage of said video images in the vertical direction thereof.
Further, the linear boundary may separate said video images from adjacent blank areas having information or decorative material superposed thereon.
It is further preferred that said means for detecting characteristics of said images includes storage means storing information relating to a plurality of different editing processes to which said video images may be subjected, and means for determining, by analysis of characteristics of the video images, the nature of an editing process imposed upon said video images.
Preferably again, said means for detecting characteristics of said images includes means for determining a dimensional extent of said editing.
Apparatus in accordance with the invention may usefully further comprise means for recognizing coded information received with said video images and indicative of at least one feature of an editing process applied to said video
<Desc/Clms Page number 3>
images. Such apparatus may further comprise means for countering said editing process to restore said video image towards its original form and content.
Apparatus in accordance with a preferred form of the invention is configured to utilize an empirical operation in the restoration process. This provides a robust apparatus of generic application.
In order that the invention may be clearly described and readily carried into effect, one embodiment thereof will now be described, by way of example only, with reference to the accompanying drawings, of which: Figure 1 shows, in flow diagrammatic form, a procedure implemented utilizing an example of the invention, and Figures 2 to 5 show in graphical form representations of video images at different points in the procedure of Figure 1.
Before referring specifically to the drawings, some general discussion relating to the nature of shrinkage, as used to alter an original video image, is presented.
In general, as mentioned previously, the technique of shrinking in video editing is used to process video images so that they become narrow and/or thin, providing blank areas above, below and/or to either side of the shrunken image. It will be appreciated that the visual content of video images so processed will be distorted, at least to an extent, but the information content of the image will in general be retained and discernable. Indeed, it is known that the human eye is quite tolerant of certain types of distortion in viewed images, and the extent and form of the alteration imposed on the images will in general be
<Desc/Clms Page number 4>
configured so as to take advantage of this, and knowledge of the types of alteration that are most likely to be used is employed, as will be described later, to ease recognition of the type of processing employed to alter a video image.
The blank areas released by shrinking the original image may be used, for example, to contain textual data, such as sub-titles or explanatory information, graphic designs intended to complement the video content of the image, or other material intended to relate to, and/or enhance in some way the original content of the video image, or a viewer's enjoyment or appreciation thereof. In any event, the boundary between the shrunken image and the blank area is usually a straight line which, in most cases, is either vertical or horizontal.
This example of the invention therefore aims to characterize the type of alteration imposed upon the video image by utilizing the technique of edge detection, and it will be appreciated that the edges to be detected are usually about the same length as the width or height of the video image itself for one directional shrinkage. For two directional shrinkage, i. e. shrinking the same image vertically and horizontally, the edges of the shrunken image and the blank area are measured to the intersection point of the vertical edge and horizontal edge. Vertical masking filters, such as vertical Sobel, can be used to detect vertical lines, and horizontal masking filters, such as horizontal Sobel, can be used to detect horizontal lines.
On the foregoing basis, vertical and horizontal edge detection can be achieved simply, utilizing the following equations:
<Desc/Clms Page number 5>
E (X, (x, y + I) - I (x, y) I, if 1 l (x, y + 1) - l (x, y) I > Te Equation 1 () =1 Equation 1 E, (X, Y) = I I (X + 1, Y)-I (X, Y) 1,'f I I (X + 1, Y)-I (X, Y) 1 > Te Equation 2 t u otherwlse f (+l)-7 () !, t7 (+l)-7 ()) > r, v (x, y) - 0, otherwise 0, orwMg
Where Te is a threshold value chosen to distinguish normal picture content differentials from the differentials that exist between picture content and blank areas ; thereby identifying picture elements (pixels) that are adjacent an edge of the video image.
For colour video images, more robust vertical and horizontal edge detection can be achieved by combining colour edge pixels detected using three prime colour signals of the images, i. e. R (x, y), G (x, y) and B (x, y). The R, G and B edge pixels can be obtained by using Equations 1 and 2 for R (x, y), G (x, y) and B (x, y) signals rather than intensity signal I (x, y).
If the number of edge pixels in a vertical line is about equal to the height of the video image, this line is considered to represent a vertical boundary between a shrunken video image and an adjacent blank area. Likewise, if the number of edge pixels in a horizontal line is about equal to the width of the video image, this line is considered to represent a horizontal boundary between a shrunken video image and an adjacent blank area. For two directional shrinkage, two lines, one vertical and one horizontal, should be detected at the same image so as to represent a boundary between a shrunken video image and an adjacent blank area.
<Desc/Clms Page number 6>
Once the boundary line has been detected, the position of the line relative to the image can be determined, thereby enabling the scale of shrinkage to be readily established. Once this has been done, the received image can be expanded by the appropriate amount to restore its size to (substantially at least) that of the original; the important issue being, of course, to render the watermark code recognizable again by the decoding circuits.
In this connection, it is possible, if the specific technique that has been used to effect the shrinkage is known, to effectively undo the shrinkage, thereby restoring the image to its original dimensions and importantly restoring the watermark code to a recognizable form. In this respect, it is feasible to have a number of shrinkage procedures listed in a memory, from which the appropriate restoring formula can be derived. This however requires, in general, the co-operation of the party transmitting the edited images in transmitting, along with the images, a code or some other data that identifies the shrinkage technique used, and it is more often the case that the specific shrinkage procedure will not be known at the output side of the transmission channel.
A generic approach to restoration of the image size is, however, derivable on an empirical basis, as will be described, so that even if the actual shrinkage procedure is not known, the generic restoring process will effect restoration with sufficient accuracy to render the watermark coding recognizable.
Suppose for example that it is determined that the vertical scale of a received image has been reduced by one third.
<Desc/Clms Page number 7>
If Io (x, y) is the intensity function of an original image at pixel (x, y), then Io (x, y + 1) and Io (x, y + 2) can be regarded as the intensities of the next two lines.
Similarly, with regard to the shrunken image, Is (x, y) is the intensity at pixel (x, y) and Is (x, y + 1) is the intensity at the corresponding pixel in the next line.
One procedure that might have been used to obtain the intensities of the shrunken image is as follows :
7, () = (/, () +7, (+1))/2 1 /+l) = (7y+l) +7+2))/2j Equation
Empirically, therefore, in order to restore the shrunken image to the dimensions of the original image, it is necessary to expand every two lines of the shrunken image to three lines. Using similar nomenclature to the above, but using Ir to indicate intensity values in the restored
image, we have :
Ir =/) Ir (x, +l) = (Is (x, +/, (x+l))/2- Equation 4 Ir (x, y+2) = , (x, y+l)
The restored image may differ somewhat from the original, but information contained within the watermark coding remains and can now be recognized.
Referring specifically to Figure 1, after starting the decoder at 1, the received image is assessed at 3 to ascertain whether it has been edited after watermark encoding. This procedure may be quite complex, because various types of editing might be encountered, but clearly a first check is to ascertain whether a watermark code is detected. If it is suspected that a watermark code is not
<Desc/Clms Page number 8>
being detected in an image that should contain one, specific editing processes are sought at 4. As mentioned above, straightforward shrinkage of images usually leaves straight-lined edges which can be detected and used as a basis for restoration. In general (unless the image sender indicates that a particular editing process has been used, in which case the appropriate restoration process can be implemented directly) the procedure carried out at 4 is to look for obvious things like the aforementioned straight edges first and, if none is found, to cycle through other stored examples of editing possibilities until a type of editing is identified. This procedure is followed at 5 by detection of the scale of editing, e. g. the amount of any size reduction brought about by shrinkage editing. The next step, at 6, is to restore the image, as nearly as can readily be achieved, to its original format, thereby permitting the decoding process to be carried out at 7 on the restored image, where the possibility of recognizing watermarked codes is significantly enhanced as compared with attempting to decode watermark data from the edited image.
Decoding having been effected, the next image can be restored for decoding or the operation discontinued as appropriate.
If multiple editing processes have been used in a sequence of video images, the complete sequence of operations shown in Figure 1 is repeated for each image in the sequence, but logically the system at 4 will first anticipate that a common editing process has been employed, and will operate on that principle until it fails to detect a watermark code. Only then will the system look for other editing processes.
<Desc/Clms Page number 9>
As an example, to illustrate the use of the invention in relation to shrunken images, Figure 2 shows an original image 10 with embedded watermark codes. Figure 3 shows at 10s this image shrunken by one third, so that there is a blank band 11 at the bottom which can, for example, contain subtitles. The straight-line boundary 12 between the (shrunken) image 10s and the blank band 11 is detected by means of a horizontal edge detector using an edge-enhanced version of the image as shown at 10e in Figure 4, and the existence of this boundary characterizes the editing as vertical shrinkage of the image. From the position of the straight-line boundary 12 in relation to the top and/or bottom of the image, the scale of shrinkage can be detected, and the image is restored as nearly as practicable to its original dimensions, as shown at lOr in Figure 5. The restored image lOr is decoded to obtain watermarked data.

Claims (12)

  1. CLAIMS 1. Recognition apparatus for recognizing watermark coding impressed upon video images received by way of a communication channel and subject to alteration prior to and/or during transfer through said channel, including means for detecting characteristics of said images indicative of at least one feature of said alteration, and means for utilizing said detected characteristics to operate upon the video images received by way of said channel in a sense tending to restore said video images toward their unaltered form prior to their application to said recognition means.
  2. 2. Apparatus according to claim 1 wherein said means for detecting characteristics is responsive to linear boundaries between said video images and their surroundings.
  3. 3. Apparatus according to claim 2 wherein said linear boundaries run horizontally and are indicative of shrinkage of said video images in the vertical direction thereof.
  4. 4. Apparatus according to claim 2 wherein said linear boundaries run vertically and are indicative of shrinkage of said video images in the horizontal direction thereof.
  5. 5. Apparatus according to claim 2 wherein said linear boundaries run vertically and horizontally and are indicative of shrinkage of said video images in both vertical and horizontal directions thereof.
    <Desc/Clms Page number 11>
  6. 6. Apparatus according to any of claims 2 to 5 wherein the linear boundary separates said video images from adjacent blank areas having information or decorative material superposed thereon.
  7. 7. Apparatus according to any preceding claim wherein said means for detecting characteristics of said images includes storage means storing information relating to a plurality of different editing processes to which said video images may be subjected, and means for determining, by analysis of characteristics of the video images, the nature of an editing process imposed upon said video images.
  8. 8. Apparatus according to any preceding claim wherein said means for detecting characteristics of said images includes means for determining a dimensional extent of said editing.
  9. 9. Apparatus according to any preceding claim further comprising means for recognizing coded information received with said video images and indicative of at least one feature of an editing process applied to said video images.
  10. 10. Apparatus according to claim 9 further comprising means for countering said editing process to restore said video image towards its original form and content.
  11. 11. Apparatus according to any preceding claim configured to utilize an empirical operation in the restoration process.
  12. 12. Recognition apparatus substantially as herein described with reference to the accompanying drawings and/or as shown therein
GB0108427A 2001-04-03 2001-04-03 Restoring compromised image data Withdrawn GB2374478A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
GB0108427A GB2374478A (en) 2001-04-03 2001-04-03 Restoring compromised image data

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
GB0108427A GB2374478A (en) 2001-04-03 2001-04-03 Restoring compromised image data

Publications (2)

Publication Number Publication Date
GB0108427D0 GB0108427D0 (en) 2001-05-23
GB2374478A true GB2374478A (en) 2002-10-16

Family

ID=9912228

Family Applications (1)

Application Number Title Priority Date Filing Date
GB0108427A Withdrawn GB2374478A (en) 2001-04-03 2001-04-03 Restoring compromised image data

Country Status (1)

Country Link
GB (1) GB2374478A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7764790B2 (en) * 1999-08-06 2010-07-27 Rovi Solutions Corporation Scaling independent technique for watermarking images

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO1999052271A1 (en) * 1998-04-02 1999-10-14 Moskowitz Scott A Multiple transform utilization and applications for secure digital watermarking

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO1999052271A1 (en) * 1998-04-02 1999-10-14 Moskowitz Scott A Multiple transform utilization and applications for secure digital watermarking

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
Recovering watermarks QIBIN Proc of SPIE 25-27 January 1999 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7764790B2 (en) * 1999-08-06 2010-07-27 Rovi Solutions Corporation Scaling independent technique for watermarking images
US8014524B2 (en) 1999-08-06 2011-09-06 Rovi Solutions Corporation Scaling independent technique for watermarking images with recorder shut-off

Also Published As

Publication number Publication date
GB0108427D0 (en) 2001-05-23

Similar Documents

Publication Publication Date Title
AU758814B2 (en) A method for inserting and detecting electronic watermark data into a digital image and a device for the same
KR101361591B1 (en) Covert and robust mark for media identification
US6714683B1 (en) Wavelet based feature modulation watermarks and related applications
US9311687B2 (en) Reducing watermark perceptibility and extending detection distortion tolerances
US5835639A (en) Method for detecting rotation and magnification in images
US6546113B1 (en) Method and apparatus for video watermarking
US7085396B2 (en) Embedding data in material
KR100602398B1 (en) Embedding and detecting a watermark in images
US6456724B1 (en) Electronic watermarking system capable of providing image data with high secrecy
US20060269098A1 (en) Image processing apparatus, image processing method, medium, code reading apparatus, and program
GB2383220A (en) Adaptive strength watermarking
EP1028585A1 (en) A method for inserting and detecting electronic watermark data into a digital image and a device for the same
KR100775774B1 (en) Embedding and detecting a watermark in an information signal
Wu et al. Attacks on digital watermarks
GB2374478A (en) Restoring compromised image data
EP1616297A2 (en) Method of detecting watermarks
Linnartz et al. Detecting electronic watermarks in digital video
EP1673724B1 (en) Detection of a watermark in a digital signal
JP4576229B2 (en) Embed Image Authentication Sign
EP1775679A1 (en) Embedding and detecting a watermark in an audio or video bit stream
Tokar et al. Digital watermarking of uncompressed video in spatial domain
Winne et al. Spatial digital watermark for MPEG-2 video authentication and tamper detection
GB2378602A (en) Applying two-part codes to stationary images
Tsang et al. Review on attacks, problems, and weaknesses of digital watermarking and the pixel reallocation attack
Jung Block-based authentication scheme for forgery attacks on digital images

Legal Events

Date Code Title Description
WAP Application withdrawn, taken to be withdrawn or refused ** after publication under section 16(1)