GB2383489A

GB2383489A - Watermarking using forced intra refresh macro-blocks

Info

Publication number: GB2383489A
Application number: GB0130672A
Authority: GB
Inventors: Paola Marcella Hobson; Anthony Richard May; Kevin Michael Mckoen
Original assignee: Motorola Inc
Current assignee: Motorola Solutions Inc
Priority date: 2001-12-21
Filing date: 2001-12-21
Publication date: 2003-06-25
Anticipated expiration: 2021-12-21
Also published as: GB0130672D0; AU2002358146A8; WO2003054798A3; GB2383489B; AU2002358146A1; WO2003054798A2

Abstract

Methods for embedding and recovering a secure watermark in an encoded video sequence. The video sequence includes intra-coded frames interspersed by predicted frames, wherein the video sequence uses at least one intermittent forced intra refresh block within a sequence of predicted frames. The method includes the step of assigning a meaning (150, 180) to said at least one intermittent forced intra refresh macro block such that said forced intra refresh macro block meaning acts as a watermark. The watermark meaning may relate to the advancing or retarding of an expected intra refresh macro block. The method may also include a step of resolving an ambiguity caused by a naturally occuring intra macro-block.

Description

Video/Image Communication And Watermarking Field of the Invention This invention relates to video transmission systems and related video encoding/decoding techniques. The invention is applicable to, but not limited to, a video compression system employing video watermarking where any tampering of a video image or portion of video image is to be detected.

Background of the Invention The ability to transmit real-time video and/or image data is a desirable characteristic of many current wireline and wireless communication systems. However, it is known that individual images/pictures, or a series of images say, in a transmitted video stream, may be subjected to 'attacks', i. e. the images may have been tampered with.

Therefore, a need exists to protect image or video transmissions from such undesirable tampering. One known technique employed to protect still/video images or documents is by the use of"watermarks".

In the context of the present invention, the terms 'video'and'image'are used interchangeably, with the term'video'generally used to represent one or more still images.

Wolfgang R, Podilchuk C, Delp E"Perceptual watermarks for digital images and video", SPIE Conference on

Security and Watermarking of Multimedia Content, January 1999, describes some state of the art watermarking methods for use with video and images.

Protection of digital media (including image and video) has also become a key standardisation topic within the multimedia industry over the last year. Police users have formally stated that they do not envisage using digitally transmitted and processed images for evidential purposes without the existence of reliable tamper detection methods.

The European Broadcasting Union has issued a second call for systems that offer watermarking of multimedia transmissions for entertainment applications. In addition, the International Standards Organisation (ISO) has set up a working group known as MPEG-21, whose essential function is to investigate digital rights management including the authentication of multimedia data.

In image watermarking, a known binary pattern or signature is embedded into an image at the moment of image acquisition. This is a well-known technique for copyright purposes. The watermark is a binary string of l's and 0's, which may represent possibly a company logo, a user authentication code, some date/time/location information, etc. Such watermarks are termed"robust", because they are designed to remain intact regardless of any post-processing of the image such as filtering, cropping, etc.

While such watermarks do provide a useful degree of protection, they cannot be wholly relied on, at present, in a court of law. The purpose of these watermarking methods is such that they are not designed to possess the required degree of surety that an image has not been tampered with, in order for the image to be used as evidence.

Furthermore, these current techniques tend to distort the image or video by treating the watermark as low-level image noise. Police users in particular are concerned that the intrusive nature of existing watermarking methods would render the images unacceptable in a court of law because it could be argued that the image has been tampered with when the watermark is embedded. In addition, watermarks inserted by many current techniques are not robust to compression, which is applied when images are transmitted/stored, or add visible distortion to the image.

Thus, there exists a need in the field of the present invention to provide a method of embedding a watermark in an encoded sequence, wherein the abovementioned disadvantages may be alleviated.

Summary of Invention In accordance with a first aspect of the present invention, there is provided a method for embedding or

recovering a secure watermark in an encoded video sequence, as claimed in claim 1.

In accordance with a second aspect of the present invention, there is provided a video communication unit, as claimed in claim 9.

In accordance with a third aspect of the present invention, there is provided a video communication system, as claimed in claim 11.

In accordance with a fourth aspect of the present invention, there is provided a method for detecting tampering of a video sequence, as claimed in claim 8.

Further aspects of the present invention are as claimed in the dependent claims.

In summary, the inventive concepts of the present invention allow a user to associate a watermark with a video sequence, without generating any visual distortion.

The watermark is implicit in the video compression method, so will be independent of the amount of compression applied. Rather than modify the image data per se, a specific mode of operation is forced onto the video codec, which will be invisible to the user, but which will allow a watermark to be carried by the video sequence.

Brief Description of the Drawings Exemplary embodiments of the present invention will now be described, with reference to the accompanying drawings, in which: FIG. 1 illustrates a flowchart of a watermark recovery method, in accordance with the preferred embodiment of the invention.

FIG. 2 illustrates a flowchart of a modified decoder to allow for naturally occurring forced intra-coded MBs, in accordance with an enhanced embodiment of the invention.

FIG. 3 shows a multiple-level bit-embedding arrangement, in accordance with the preferred embodiment of the invention.

Description of Preferred Embodiments In video coding systems using ITU H. 263 [ITU-T Recommendation, H. 263,"Video Coding for Low Bit Rate Communication"] video compression technology, the image data are compressed in macro blocks (MBs). These MBs comprise four luma and two chroma blocks of 8x8 pixels.

The first frame of the video sequence is transmitted as intra-coded information. The intra-coded information is followed by faster inter-coded information.

Clearly, in most commercial video systems there exists a bandwidth constraint. In particular, bandwidth is both a

critical and valuable commodity in low bandwidth systems, where the intra-coded frame frequently contains a relatively large amount of data. Therefore, a video transmission that incorporates intra-coded frames takes a relatively long time to transmit.

The first image transmitted in a video sequence is always an intra-coded (I) frame. The I-frame contains data relating to the whole of the image. Due to the large amount of data required in the I-frame, the video system designer often faces a compromise between the spatial quality of the image and the transmission time. This is particularly true for low-bandwidth systems, such as the TErrestrial Trunked RAdio system (TETRA) as defined by the European Telecommunications Standards Institute (ETSI).

Subsequent frames to the I-frame (s) are therefore encoded as predicted (P) frames. The use of prediction with subsequent frames improves the frame rate. P-frames use the temporal similarity between the last transmitted frame and the current frame to reduce the amount of data to be transmitted.

In the context of the present invention, and the indications of the advantages of the present invention over the known art, the expression"video transmission", as used in the remaining description, encompasses various video techniques. These typically involve real-time, or near real-time, transmission of images at various data rates. The expression of"video"in the subsequent

description also encompasses image transmission, which is generally viewed as a one-frame video. Furthermore, the video techniques concerned may also include, for example, video that is streamed, or encoded for storage with the ability for the video images to be viewed later.

Within a P-frame, some MBs may not be encoded at all (skipped macro blocks). However, the H. 263 standard also specifies that a certain number of MBs must be intracoded per frame, to reduce errors resulting from the accumulation of prediction errors (forced intra refresh).

The inventive concept of the present invention allows a watermark to be carried by an H. 263 compressed video stream by modifying the operation of the codec such that a meaning can be assigned to the forced intra refreshed MBs.

First, the watermark must be represented as a series of binary digits. In its simplest embodiment, the method allows insertion of one bit per forced intra refresh MB of the video. Assuming a forced intra macro block is sent at least every frame, this allows a minimum of one bit per frame to be sent.

The user must specify in advance the forced intra refresh sequence for the codec, which must be known by the decoder. In order to embed a watermark bit'1', the specified intra refresh block number should be advanced by a given step size, for example'I'or'2'. To embed a watermark bit'0', the specified intra refresh block

number should be retarded by a given step size, for example'0','1'or'2', which may or may not be the same step size which was used to advance the block number when a bit'1'is embedded. Note that the arithmetic is modulo the number of MBs in the image.

As an example, let the MB forced intra refresh sequence for a quarter common intermediate format (QCIF) image containing 99 MBs (as known to those skilled in the art) be those MB numbers shown in Table 1. Eighteen frames are shown and it is assumed that only one intra MB is forced per frame.

Table 1: Example MB forced intra refresh sequence.

Frame 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 MB No 45 30 1 78 66 12 88 99 11 2 43 38 55 8 18 94 29 70 Let the watermark to be embedded advance the sequence number by a step size of'2'for a bit 11'and retard it by'2'for a bit'0'. Table 2 shows the changes to the MB forced intra refresh sequence.

Table 2: Example of applying a watermark to the MB forced intra refresh sequence.

Frame 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 MB No 45 30 1 78 66 12 88 99 11 2 43 38 55 8 18 94 29 70 Water- 0 0 0 1 0 1 0 1 1 0 0 0 1 0 1 1 0 1 mark bit New MB 43 28 98 80 64 14 86 2 13 99 41 36 57 6 20 96 27 72 No

As an example, consider a QCIF image, in which there are 99 macro blocks. Each MB must be updated at least once per 132 frames. The particular MB or MBs to intra refresh in each frame is/are chosen based on a pseudorandom sequence, which is made known to the decoder in advance. During encoding of each frame, the position of the intra-refreshed block is advanced or retarded depending on the watermark bit value. This will override any attempts by the codec to skip those particular MBs, if appropriate.

At the decoder, the expected positions of forced intra blocks are sought. The watermark bit values are determined from examining the MBs at a distance of the step size from the expected MB to find where a forced intra has occurred.

Referring now to FIG. 1 a flowchart of the watermark recovery process 100 for, say, a step size = S, is shown in accordance with the preferred embodiment of the present invention. An incoming bit stream 110 is received at the decoder. The decoder has previously arranged with the encoder a list of stored values of forced intra refresh MB sequence, as shown in step 130.

The received actual forced intra MB location (s) in the bit-stream is/are then compared to these previously stored values of forced intra refresh MB sequence, for each encoded frame, as in step 120.

The comparison step then provides an indication of the value of the watermark bit or an indication of whether tampering of the video bit-stream may have occurred. If the location of the forced intra MB (s) are in the expected location, plus an advance of IS' (the step size) in step 140, it is assumed that the watermark bit was a '1', as shown in step 150.

If the location (s) of the forced intra MB (s) is/are not in the expected location (s) plus an advance of IS'in step 140, a determination is made as to whether the location (s) of the forced intra MB (s) is/are in the expected location (s), minus a retard of'8', as shown in step 160. If the location (s) of the forced intra MB (s) is/are in the expected location (s), minus a retard of ils', in step 160 it is assumed that the watermark bit was a'0', as shown in step 170.

If the location (s) of the forced intra MB (s) is/are not in the expected location (s), minus a retard of'8', as in step 160, it is assumed that no recovery is possible and that image tampering may have occurred, as shown in step 180.

Advanced Recovery Process It is possible that during normal compression of the video, certain MBs may be intra-coded as a result of scene content, such as highly variable image content, uncovered background, or other places where motion estimation is very inaccurate. These would therefore be

'naturally occuring'intra MBs, rather than MBs forced in accordance with the watermarking scheme of the present ainvention. It could occur that in addition to the forced intra refresh that is imposed at MB N+S (MB number N plus step of S), a naturally occurring intra MB is required at MB N-S. Hence, it is possible that the decoder would not know if this should be a watermark bit Ill or watermark bit'0'.

Hence, to avoid this problem, it is within the contemplation of the invention that when embedding a watermark bit'1', the encoder would preferably examine the MB N-S value when forcing an intra refresh on MB N+S.

If MB N-S is also intra coded, then in addition to the forced intra MB and the natural intra MB, the encoder continues and will force an intra refresh at the next MB+S in the sequence.

Similarly, when embedding a watermark bit'0', the encoder would preferably examine the MB N+S when forcing an intra refresh on MB N-S. If MB N+S is found to be intra coded, then in addition to the forced intra MB and the natural intra MB, the encoder continues and will force an intra refresh at the next MB-S in the sequence.

This process repeats until there is no ambiguity.

In a similar manner, the decoder operation is preferably adapted to manage the enhanced recovery process 200, as shown in FIG. 2. The same process applies, as described with respect to FIG. 1, apart from the determination in step 140.

If the location of the forced intra MB (s) are in the expected location, plus an advance of'S' (the step size) in step 140, it is not automatically assumed that the watermark bit was a'1'. In the enhanced watermarking mode, a determination is made as to whether a naturally occurring intra MB (s) has been effected. This occurs as follows: If a forced intra MB is found in the expected

location, plus an advance of'8', a determination is made as to whether a naturally occurring intra MB is in the expected location, minus a retard of'8', as shown in step 280. If no naturally occurring intra MB is found in the expected location, minus a retard of'S', in step 280, then it is assumed that the watermark bit was a Ill, as shown in step 285.

However, if it is determined that a naturally occurring

intra MB is in the expected location, minus a retard of 'S', in step 280, it is assumed that an ambiguity has occurred, and the decoder will move to the next expected location in step 290, before making a watermark determination. Such a decoding process repeats until it is determined that an ambiguity is not found.

It is noteworthy that FIG. 1 and FIG. 2 both contain a state (step 180 and step 270 respectively) where no recovery of the watermark bit is possible. This may arise where a forced intra MB is not found at either the expected location +S or the expected location-S. An interpretation that some image corruption or intentional tampering may be made. However, it would not be possible

to determine by this method alone if this is due to loss of MBs or complete frames of the video.

For H. 263 images each macro block must be updated at least once every 132 times that coefficients are transmitted for a macro block. For a common intermediate format (CIF) image, which contains 396 macro blocks, this requires forcing three intra macro blocks per compressed frame. This would allow three watermarking bits per frame to be encoded. For QCIF images, which only contain 99 macro blocks, only 0.75 bits can be encoded per frame.

To minimise complexity, most implementations of H. 263 QCIF encoders force one intra macro block per compressed frame, which would allow one watermarking bit to be encoded, in accordance with the preferred embodiment of the invention.

It is within the contemplation of the invention that, in order to improve efficiency, the encoder and decoder may arrange for more than the minimum number of forced intra MBs per frame. This would increase the number of watermark bits that can be embedded per frame. This extension to the method described above with respect to FIG. 1 and FIG. 2 remains compliant with the H. 263 standard.

Multiple Bit Insertion In accordance with a further embodiment of the present invention, multiple bit insertion is used. Multiple bit insertion is possible if the macro block update sequence

number is treated as a variable, which is allowed to take on any integer value (as in an amplitude shift keying modulation scheme), modulo the number of MBs in a frame.

An example of such an embodiment 300 is shown in FIG. 3, with 2 bits, and thereby 4 (watermark bits) shift values 350 allowed, 00' 360, 01' 365, 10' 370 and 11' 375.

Each watermark bit reflects an increase or retarding of an MB sequence number 310, between'-M'and'M'as shown. For example, if the encoder and decoder settle on a value of M=4, then assuming'01'was to be embedded, the skipped MB position would be retarded by'2'.

It is within the contemplation of the invention that the step sizes do not necessarily have to be equal, and other changes to the MB number are envisaged in order to force an intra refresh. One such example is shown in Table 3 where three bits can be encoded per MB.

Table 3: Example of a three bit embedding scheme

Bits to embed Advance/retard by 000 0 001 +2 010 +5 011 +9 100 -1 101 -3 110-5 dz

H. 263 Compliance Advantageously, the inventive concepts of the present invention can be adapted to ensure compliance with the H. 263 standard. Such compliance is difficult, as H. 263 requires a forced update of each macro block at least every 132 times that coefficients are transmitted for a macro block. Forced updates are required due to errors in the picture continuing to propagate when only predicted MBs are sent.

Transmission errors can cause significant degradation of image quality if MBs are not updated. As an example, a MB sequence number 22 may be modulated to become 23 if the watermark bit'1'is to be embedded (step S=l in this case). It is possible that within the 132 frames, MB sequence number 21 would be modulated to become 22 when it reaches its turn for modulation. However, it is equally likely that a'a'is to be embedded, and MB 21 becomes 20 or whatever the sequence number step-size results in.

Over a long period, all MBs would eventually be updated.

However, such an arrangement would no longer be H. 263 compliant, and reduced image quality may persist until an update occurs.

Hence, to guarantee compliance, the refresh status of all macro blocks is tracked. If a macro block has not been updated after 131 frames, then the macro block is forced in the next frame, before the random sequence and

watermarking algorithms are applied to the encoded frame.

The random nature of the forced intra macro block algorithm would ensure that only a few macro blocks would have to be forced in this way.

If statistical analysis of the technique shows that typically, for example, ten intra macro blocks are required to be updated in the same frame, then the compliance algorithm can be modified so that macro blocks are forced if they have not been updated for between, for example, 120 and 131 frames. This would ensure that only one or two macro blocks are updated per frame. Such an approach would minimise the chance of the technique adding too much additional data to any one frame.

A yet further embodiment of the present invention, to ensure compliance with the H. 263 standard, is to embed the watermark twice sequentially in the encoded bitstream-once as normal, and once inverted. The inversion ensures that any MB position in the intra refresh sequence, which was changed by the step size S when the normal watermark was embedded, will not be changed when the inverted watermark is embedded.

The change to the basic method is illustrated in Table 4, with step size S=l to embed a watermark bit'1', and a step size 8=0 to embed a watermark bit'0'. It is noteworthy that repetition of the watermark increases robustness as there will be two bits available for a 'O'/'l'value decision at the decoder. Efficiency is

maintained by increasing the number of forced intra MBs per frame.

Table 4: Example of repetition of the inverted watermark to ensure H. 263 compliance.

Frame 1 2 3 4 5 ... 49 50 51 52 53 ... 99 100 MB No's 45 30 1 78 66 12 88 99 11 2 15 45 30 1 78 66 12 88 99 11 15 45 30 Water- 0 0 0 1 0 1 0 1 1 0 0 1 1 1 0 1 0 1 0 0 1 1 0 mark bit New MB 45 30 1 79 66 13 88 1 12 2 15 46 31 2 78 67 12 89 99 11 16 46 30 No's

This guaranteed H. 263 compliant method saves on the complexity of tracking refreshed macro blocks. However, it does require that the minimum number of forced intra MBs per frame be increased, and that the step size for either a watermark bit'1'or watermark bit 0'be 0'.

In this enhancement to the basic method, when an ambiguity occurs during the embedding of the normal watermark, in addition to the forced intra MB and the natural intra MB, the encoder goes on to force an intra refresh at the next MB+/-S in the sequence. To ensure that H. 263 compliance is not lost, a record is kept of the location of the ambiguity, and a matching ambiguity is forced in the embedding of the inverted watermark.

As an example of ambiguity handling, consider the watermark 0001010110... and an MB sequence of 45 30 1 78 66 12 88 99 11 2..... Table 4 shows that in Frame 1, the new MB positions would be 45 and 30. However, if we

assume that MB 31 was a naturally occurring intra MB, then when examining MB 30 and 31, the decoder would not know if a 1'or a 0'should have been embedded.

Therefore, in addition to these two forced intras, the encoder moves to the next MB in the sequence, MB 1 in Frame 2. An intra MB is forced at MB 1, and assuming there is no ambiguity with MB 99 or MB 2, the encoder moves on to MB 79 in Frame 2.

When embedding the inverted watermark, the encoder uses the list of ambiguities to force ambiguity in the next set of frames. In this manner, the MB positions used in embedding the inverted watermark exactly match the positions used to embed the normal watermark. Continuing the example above, to embed the inverted watermark, we wish to embed 1110101001... at MB positions 45 30 1 78 66 12 88 99 11 2.... In Frame 49 of the next series of 99 MBs (for QCIF images), an intra is forced at MB 46, to embed the watermark bit'1'. An ambiguity is then created in the Frame 50, by forcing an intra at MB 30 and MB 31. The next watermark bit'1'will then be embedded by forcing an intra at MB 2 of Frame 50. Any ambiguity here will be ignored, as the correct watermark bit can be recovered from the previously embedded normal watermark.

The encoder now moves on to MB 78 in Frame 51.

If an ambiguity occurs during embedding of the inverted watermark, which was not present when the normal watermark was embedded, it is ignored. In this manner, the exact sequence matching is not lost. The decoder must recover the watermark bit from the previous

embedding of the normal watermark only, and loses the benefit of having two bits for reduced errors in watermark recovery.

In summary, all the basic and the enhanced watermarking methods described above minimise the additional overhead required to send intra macro blocks. Hence, the quality of an image at a given bit-rate is maintained.

When there are only a few natural intra macro blocks in an encoded frame, the chance of ambiguity is low, and the watermark can be inserted with only a small chance of increasing the number of bits required to encode the frame. When there are many natural intra macro blocks, the additional forced intra macro blocks that are required to deal with ambiguity, have only a small percentage impact on the total amount of data that is required to encode the frame.

Advanced Protection The aforementioned methods are secure for low sophistication attacks where an attacker decodes, modifies or copies, and then re-encodes the video sequence. A more sophisticated attacker would possibly be able to track the intra MBs during decoding, and would have means to repeat the pattern of intra MBs in the reencoded sequence. To protect against this type of attack, it is envisaged that the watermark to be embedded should be a signature or hash of the image content.

Such a signature is created at the encoder by arithmetic operations on the image content, and is unique to the content. The signature is encrypted (as known in the art) and embedded as the watermark, using the inventive techniques described above. At the decoder, the signature is regenerated and compared with the extracted watermark. If the image content has been tampered with, the regenerated signature will not match the recovered watermark.

The inventive concepts of the present invention provide a simple means of embedding a watermark in compressed video bit-stream, such as H. 263 compressed video. Such methods may be readily applied to video transmission systems, for example a TETRA packet data system as well as H. 263 communication systems. The invention is also applicable to any block based compression system (such as H. 263 and MPEG4) where an invisible watermark is needed. Although watermark tampering is not obvious from examining the received bit-stream, this property makes the method very robust to intentional attacks or attempts to remove the watermark, which is essential for multimedia security systems.

Thus, a method for embedding or recovering a secure watermark in an encoded video sequence has been described. The encoded video sequence includes intracoded frames interspersed by predicted frames, wherein the video sequence uses at least one intermittent forced intra refresh block within a sequence of predicted

frames. The method includes the step of assigning a meaning to said at least one intermittent forced intra refresh macro block such that said forced intra refresh macro block meaning acts as a watermark.

It will be understood that the method for embedding a secure watermark in an encoded video sequence described above provides for a means of embedding tamper evident information in a compressed video bit-stream in a way that does not alter the image data in any way. Thus, no loss of subjective or objective quality is incurred when this method is used. The method is compliant with H. 263 standard video compression.

Whilst the specific and preferred implementations of the embodiments of the present invention are described above, it is clear that one skilled in the art could readily apply variations and modifications of such inventive concepts.

Claims

Claims 1. A method for embedding or recovering a secure watermark in an encoded video sequence having intra-coded frames interspersed by predicted frames, wherein the video sequence uses at least one intermittent forced intra refresh block within a sequence of predicted frames, the method characterised by the step of: assigning a meaning (150, 180) to said at least one intermittent forced intra refresh macro block such that said forced intra refresh macro block meaning acts as a watermark.
2. The method for embedding or recovering a secure watermark in an encoded video sequence according to claim 1, the method further characterised by the steps of: communicating said meaning between an encoder transmitting said encoded video sequence and a decoder receiving said encoded video sequence prior to sending said video sequence, wherein said watermark meaning is a bit value that relates to an advancement or retarding of an expected intra refresh macro block.
3. The method for embedding or recovering a secure watermark in an encoded video sequence according to claim 2, the method further characterised by the steps of: checking a corresponding retarding value, if said watermark meaning is a bit value that relates to an advancement of an expected intra refresh macro block ; or checking a corresponding advancement value, if said watermark meaning is a bit value that relates to a retarding of an expected intra refresh macro block.

<Desc/Clms Page number 23>
4. The method for embedding or recovering a secure watermark in an encoded video sequence according to claim 2 or claim 3, the method further characterised by the step of: treating a macro block update sequence advancement or retarding value as a variable able to take on any integer value, to facilitate multiple watermark bit insertion into said encoded video sequence.
5. The method for embedding a secure watermark in an encoded video sequence according to any preceding claim, the method further characterised by the steps of: tracking a refresh status of substantially all macro blocks; and introducing forced intra macro blocks into the video sequence if a macro block has not been updated after a predetermined number of frames.
6. The method for embedding a secure watermark in an encoded video sequence according to any preceding claim, the method further characterised by the step of: embedding the watermark twice sequentially in the encoded bit-stream, wherein the embedding includes once as a normal watermark value and once as an inverted watermark value.

<Desc/Clms Page number 24>
7. The method for embedding a secure watermark in an encoded video sequence according to any preceding claim, the method further characterised by the step of: forcing an intra refresh at a subsequent macro block in said video sequence, in addition to any forced intra macro blocks and any natural intra macro blocks if an ambiguity occurs during the embedding process.
8. A method for detection of tampering of a video sequence, based on inability to recover any or all of the watermark bits that were embedded according to any of the preceding Claims 1 to 7.
9. A video communication unit adapted to perform the method steps of any of the preceding claims.
10. The video communication unit according to claim 9, wherein the video communication unit is, or is contained within a mobile phone, a portable or mobile radio, a personal digital assistant, a laptop computer or a wirelessly networked PC.
11. A video communication system adapted to facilitate the performance of the method of any of claims 1-8.
12. A method for recovering a secure watermark in an encoded video sequence substantially as hereinbefore described with reference to, and/or as illustrated by, FIG. 1 or FIG. 2 of the accompanying drawings.