WO2006134517A2

WO2006134517A2 - Encryption and decryption of digital color image signals

Info

Publication number: WO2006134517A2
Application number: PCT/IB2006/051816
Authority: WO
Inventors: Anil Yekkala; Elisabeth F. M. Steffens; Narendranath Udupa
Original assignee: Koninklijke Philips Electronics N.V.
Priority date: 2005-06-17
Filing date: 2006-06-07
Publication date: 2006-12-21
Also published as: WO2006134517A3

Abstract

A colour image signal is partially encrypted and decrypted. At least part of the data that represents a luminance component of the source colour image signal is encrypted and inserting in the partially encrypted digital image signal. All colour properties of pixel values from the source colour image that have luminance according to the luminance component are inserted without encryption. Preferably, the source colour image signal is copied into the partially encrypted digital image signal, except for a replacement of the luminance component by the encrypted data. Also preferably encryption is applied only a zero frequency coefficient of a Discrete Cosine transform of luminance components for the pixel locations in the block and non-zero frequency coefficients of the Discrete Cosine Transform. In a further embodiment a subset of the blocks is selected for the encryption of data derived from the zero frequency coefficients, and luminance information of the blocks outside the subset is inserted into the partially encrypted digital image signal without encryption.

Description

Encryption and decryption of digital color image signals

The invention relates to encryption and decryption of digital colour image signals.

Encryption of digital colour image signals, in particular video signals, such as

MPEG encoded video, serves to ensure that it is not possible to view a digital colour image if no decryption key is available. In principle, all of the digital data that represents a digital colour image signal may be encrypted. However, this type of encryption requires considerable computational effort both for encryption and decryption. As a result real time decryption requires a considerable amounts of secure computation circuitry. This makes full encryption economically infeasible.

Techniques that result in less computational effort have been described in an article by Iskender Agi and Li Gong, titled "An Empirical Study of Secure MPEG Video Transmissions" and published in the proceedings of the internet society symposium on Network and Distributed Systems Security, pages 137-144, San Diego CA. Agi et al describe selective encryption of I-frames and headers of MPEG encoded data.

It is also known to address this problem by encrypting only the zero frequency (DC) component of blocks of pixels that are encoded using the MPEG standard. MPEG divides each colour image into a matrix of square blocks and computes the DCT (Discrete Cosine Transform) of the pixel values in each block. The resulting DCT coefficients include a coefficient for zero frequency (also called the DC coefficient, which is proportional to the average of the pixel values in the block) and AC coefficients for non-zero frequencies that represent deviations from the average. The DC coefficient is included in the encoded MPEG signal. The AC coefficients are encoded by means of run length coding. In order to provide computational efficiency of encryption of MPEG it is known to perform selective encryption. One method of achieving this is to encrypt only the DC coefficient of each block and include the encrypted DC coefficient in the MPEG signals. Thus, each DC coefficient is replaced by an encrypted coefficient, which is a function of the original DC coefficient and an encryption key. The resulting MPEG signals can be decoded with a standard decoder, but it has been found that no useful image is obtained if the DC coefficients are not decrypted.

However, these forms of partial encryption still require considerable computational effort, which makes it necessary to provide significant secure computation circuitry, or decreased availability of the computation circuitry for other purposes

Among others it is an object of the invention to reduce the computational effort that is needed for encryption and/or decryption of digital colour image signals. According to one aspect only part of the luminance component of the image data is encrypted with a secure encryption technique. The colour properties of the pixels that have the encrypted luminance are not encrypted or encrypted with a weaker encryption technique (i.e. an encryption technique that requires less computation time and/or circuitry for encryption and/or decryption). It has been found that this provides sufficiently strong encryption to make undecrypted viewing unattractive (because the human eye is more sensitive to changes in luminance in comparison to color), but still requires less computational effort. Preferably, the colour properties are not encrypted at all.

In an embodiment a standard format, like an MPEG format is used for the partially encrypted signal. The standard format is of a type that can be used for unencrypted signals. In this embodiment partial encryption involves substituting encrypted luminance parameter values in selected parts of the signal that would contain unencrypted these luminance parameters in an unencrypted signal. Thus, the partially encrypted signal can be handled as a normal, unencrypted signal. Standard equipment can be used to transmit and process the partially encrypted signal, except that standard decoding of the partially decrypted signal (i.e. without decryption) yields a result in which the original source signal is not perceivable, even though the standard equipment can normally decode the signal.

In further embodiments partial encryption (and decryption) is realized by copying a standard signal of this type except for selective replacement of at least part of the luminance data by encrypted (or decrypted) data from the signal. In a preferred embodiment Discrete Cosine Transform coefficients for blocks of pixel locations are used (such as known for MPEG encoding for example) and only the zero frequency coefficients of the luminance component of at least part of the blocks are encrypted. It has been found that this is sufficient to provide effective protection against viewing the source data images without decryption. As a result only a small amount of computation is needed to perform encryption and decryption.

In a further embodiment the zero frequency coefficients of the luminance component of only part of the blocks is encrypted. It has been found that even this is sufficient to provide effective protection against viewing the source data images without decryption. Preferably data is added to the partially encrypted signal to signal for which of the blocks the zero frequency coefficients of the luminance component is encrypted. In this way decryption can be controlled flexibly. In another embodiment the number of blocks for which frequency coefficients of the luminance component are encrypted is selected adaptively, by adapting this number until a computed measure of deviation between the original image and a result of decoding the partially encrypted signal without decryption equals at least a threshold. In this way the amount of computation needed for encryption and decryption can be minimized, while maintaining an effective amount of decryption.

These and other objects and advantageous aspects will be illustrated by means of non- limitative examples using a description of the following figures.

Figure 1 shows an encryption apparatus;

Figure Ia shows an alternative encryption apparatus; Figure 2 shows a decryption apparatus;

Figure 3 shows a flow chart of encryption;

Figure 4 shows a flow chart of decryption;

Figure 5 shows a decryption apparatus;

Figure 6 shows an encryption apparatus with a feedback loop.

Figure 1 shows functional elements of an encryption apparatus. The encryption apparatus has an input 10 for receiving video information, a luminance encoding circuit 12, a first and second colour encoding circuit 14a,b, an encryption circuit 16 and a multiplexer 18. Luminance encoding circuit 12 and first and second colour encoding circuits 14a,b have inputs coupled to input 10. The output of luminance encoding circuit 12 is coupled to multiplexer 18 via encryption circuit 16, which has an encryption key input EK. First and second colour encoding circuit 14a,b are coupled to multiplexer 18 without intervening encryption circuit. Multiplexer 18 has a further input 11 for further data relating to the video information and an output 19 for an encrypted and encoded video signal.

It should be emphasized that figure 1 shows functional elements, which means that in practice the various functions may be combined in different ways. Figure Ia shows such an alternative combination. Herein an MPEG encoder 100 is provided, a luminance information detector 102 and a switching multiplexer 104. Luminance information detector 102 has an input coupled to an output of MPEG encoder 100. Luminance information detector 102 controls switching multiplexer 104. Switching multiplexer 104 has a first input coupled to the output of MPEG encoder 100 via encryption circuit 16 and a second input coupled to the output of MPEG encoder 100 without intermediate encryption circuit 16. In operation luminance information detector 102 detects when MPEG encoder 100 outputs specific luminance information and control switching multiplexer 104 to pass MPEG encoded data after encryption selectively when MPEG encoder 100 outputs the specific luminance information and without encryption otherwise. Figure 2 shows functional elements of a decryption apparatus. The decryption apparatus has an input 20 for receiving encrypted and encoded video information, a demultiplexer 22, a decryption circuit 24, a luminance decoding circuit 26, a first and second colour decoding circuit 27a,b, an image signal generator 28 and a display screen 29. Input 20 is coupled to demultiplexer 22 which has a first output coupled to luminance decoding circuit 26 via decryption circuit 24 which has a decryption key input DK. Demultiplexer 22 which has second and third outputs coupled to first and second colour decoding circuits 27a,b, without intervening decryption circuits. Luminance decoding circuit 26, and first and second colour decoding circuit 27a,b have outputs coupled to image signal generator 28, which in turn has an output coupled to display screen 29. In an alternative embodiment part or all of the functions may be performed by a programmable computer that has been programmed with a program for this purpose.

Typically a decryption apparatus receives the output signal of the encryption apparatus after it has been transmitted via a transmission channel (e.g. via a cable TV network and/or wireless broadcast) and or stored in a storage devices such as on an optical or magnetic disk, a tape or in a semi-conductor memory.

It should once more be emphasized that figure 2 shows functional elements, which means that in practice the various functions may be combined in different ways. For example, the inverse of the embodiment of figure Ia may be used, with a switching multiplexer that passes data received after decryption or without decryption dependent on whether a detector detects that specific luminance information has been received. The switching multiplexer is followed by an MPEG decoder in this case. In an alternative embodiment part or all of the functions may be performed by a programmable computer that has been programmed with a program for this purpose. In operation encryption apparatus receives a digital source image signal for example from a camera (not shown), possibly via some transmission system. Camera images are encoded, for example according to an MPEG standard that is known per se. Part or all of the encoding may be performed outside the encryption apparatus, or the encoding may be performed entirely in the encryption apparatus. Encoding includes subdividing an image into blocks (e.g. of 8x8 pixel locations) and computing Discrete Cosine Transforms (DCT) of the luminance and colour values for the pixel locations in each block. The resulting DCT coefficients are used to form the encoded signal, typically after quantization of coefficients and run length encoding of part of the coefficients. In the apparatus of figure 1, part of the coefficients are also encrypted. Figure 3 shows a flow chart that illustrates encryption. In a first step 31 a block of 8x8 pixel locations is selected. In a second step 32, a DCT is applied to luminance components Y of colour pixel values for the pixel locations in the block (e.g. by luminance encoding circuit 12). In a third step 33 the resulting zero frequency DCT coefficient is encrypted (e.g. by encryption circuit 16), after quantization, if quantization is used, and inserted into an encoded and partially encrypted signal (e.g. by multiplexer 18) together with information that describes DCT coefficients for other frequencies, the latter without encryption. In a fourth and fifth step 34, 35 DCT's are applied to the UV components of colour pixel values for the pixel locations in the block and resulting coefficients are used to derive data that is inserted into an encoded and encrypted signal. No encryption step equivalent to third step 33 is performed for the UV components before the insertion. After fifth step 35 the process returns to first step 31 for handling a next block.

Typically output 19 is coupled to a transmission medium, such as cables of a cable TV system via a cable head end (not shown), or to a wireless broadcast medium via a transmitter (not shown). Alternatively output may be coupled to store copies of the data on a storage medium, such as an optical disk for distribution. The encoded and partially encrypted signal that is produced by the steps of figure 3 is transmitted over the transmission medium or copies of it are stored on the distribution media.

The decryption apparatus performs the inverse of the encryption. The encoded and partially encrypted signal is recovered from the transmission medium or a distribution medium such as an optical disk, partially decrypted, decoded and used to control an image displayed by display screen 29. For the decoding known MPEG decoding techniques may be used.

Figure 4 shows a flow chart that illustrates decryption. In a first step 41 information that encodes a pixel data is input, which corresponds to the encoded and partially encrypted signal that has been formed with the process of figure 3 for example. In a second step 42, the decryption apparatus detects whether the data encodes the DC component of the DCT of luminance information. If so, decryption apparatus executes a third step 43, decrypting the DC component of the luminance information. If not, the decryption apparatus skips the third step 43. Subsequently a fourth step 44 is executed, performing an inverse

DCT. In a fifth step 45 the information is used to control image content on display screen 29. It should be appreciated that the flow charts have been simplified for the sake of example. In practice many variations are possible. For example, encoding and decoding of different colour components and luminance may be performed in parallel or different parts of the process may be pipelined (a task being performed for certain data while a preceding task is already performed for the next data).

It should be appreciated that only a small amount of decryption is needed in this way. This can be performed by temporarily occupying a processing circuit that is used for other purposes at other times, or a slow special purpose decryption processor, which performs the required decryption in parallel with other parts of the decryption apparatus that perform other tasks.

An embodiment has been shown wherein a selection is made whether or not to encrypt dependent on whether the data concerned encodes a zero frequency coefficient of a DCT of luminance values for a block of pixel locations, to encrypt selectively only the zero frequencies coefficient of the DCT of the luminance information and not the remaining coefficients or any DCT coefficients of UV information. Typically, during MPEG encoding for example, the DCT coefficients are quantized. Preferably encryption is performed after quantization and decryption is performed before de-quantization.

However, it should be appreciated that other embodiments are possible. For example, instead of only encrypting the DC coefficient of the luminance component a greater number of coefficients, including the DC coefficient and one or more AC coefficients could be encrypted. Preferably these AC coefficients include coefficients for relatively lower frequencies, while coefficients for relatively higher frequencies remain unencrypted. This provides an efficient way of making undecrypted information unintelligible with a minimum of computational effort for encryption and decryption. In a further embodiment all DCT coefficients of the luminance component are encrypted the colour components remaining unencrypted. In another embodiment only one or more coefficients of the luminance components of I frame are encrypted, leaving residue information for P and B frames unencrypted. This also provides an efficient encryption scheme. Alternatively, one or more coefficients of the luminance component of only P and/or B frames but not of I frames are encrypted. However, this is less efficient in terms of reduction of intelligibility.

In another embodiment only part of the bits of the encrypted luminance component (e.g. part of the bits of the zero frequency DCT coefficient of the luminance component) are encrypted, leaving the remaining bits unencrypted. Preferably the most significant bit or bits are encrypted (including the sign bit, if any), leaving the less significant bits unencrypted. This is highly efficient in terms of reduction of intelligibility. For example, only the most significant bit (typically the sign bit) is encrypted in one embodiment. In another embodiment the zero frequency coefficient of the DCT of luminance values is encrypted and decrypted only for selected blocks. This further reduces the amount of computational effort needed for encryption and decryption. The blocks may be selected periodically (for example every Nth block, where N is an integer greater than one, N=2 or 3 for example) or some more random selection may be made.

Preferably, in the embodiment wherein the zero frequency coefficient of the DCT of luminance values is encrypted only for selected blocks, information is added to the encoded and encrypted signal that indicates the type of block selection that is used, e.g. in the form of a number N that indicates the encryption period and/or an offset number that indicates the first encrypted block in a frame. In this way encryption techniques of different strength can be used for the same decryption apparatus. Figure 5 shows a decryption apparatus that implements this technique. The apparatus comprises a selection unit 50, a luminance information detector 52, a decryption circuit 24 a multiplexer 54 and an MPEG decoder. Multiplexer 54 has a first input coupled to the input 20 of the apparatus via decryption circuit 24 and a second input coupled to the input 20 of the apparatus without intervening decryption circuit 24. An output of multiplexer 54 is coupled to an input of MPEG decoder. Luminance information detector 52 has an input coupled to apparatus input 20 and an output coupled to a control input of multiplexer 54. Selection unit has an input coupled to the apparatus input 20 and an output coupled to an enable input of luminance information detector 52. In operation selection unit 50 reads information from the input signal that encodes selection criteria of blocks, e.g. in the form of a number N that indicates the period between encrypted blocks. Selection unit 50 monitors the arrival of blocks and generates an enable signal when data for an encrypted block has arrived (e.g. each time after counting N blocks, N having been read from the input stream). Luminance information detector 52 detects whether zero frequency coefficient of the DCT of luminance values for a block is received. If so and if enabled by selection unit 50 luminance information detector 52 controls multiplexer to pass the data after decryption by decryption circuit 24 to MPEG decoder 56. Otherwise, luminance information detector 52 controls multiplexer to pass the data from input 20 without decryption.

As an alternative, information may be included in block headers to indicate whether or not decryption is needed. Instead of using periodic selection of blocks for encryption a more random selection may be used.

During encryption the average rate at which luminance data for blocks is encrypted (e.g. the value of N) may be set to any desired value. Preferably, excessively long periods between encrypted blocks is avoided, as this entails the risk that undecrypted data becomes acceptable for viewing. An average frequency corresponding to a period of N less than or equal to three is preferred.

Figure 6 shows a further embodiment wherein this frequency is selected adaptively during encryption. A comparator circuit 60 is provided that computes a measure of deviations between the input signal and signals obtained by decoding without decryption, and by using a feedback loop to decrease or increase the frequency (by controlling a multiplexer 62 to pass encrypted or unencrypted data) when the measure of deviation is above or below a desired value respectively. As the measure of deviation a standard deviation of the deviation (root mean square of the differences) may be used for example or a spectral density of the deviations at some selected frequency, or a weighted average of spectral densities or any other measure.

The frequency may be adapted for each image individually, for example downward until the measure of deviation for the image has at least a threshold value. Alternatively the frequency may be regulated on a running basis for a video signal, by reducing or increasing the value of the frequency on an ongoing basis when the measure of deviation is below or above the desired level respectively. When a slow feedback is used this may lead to incidental images that have insufficient measure of deviation, but overall the video signal will be sufficiently encrypted. Different sample frequencies may be used for luminance Y and colour UV information. The reduction of computational effort for encryption depends on the relative sampling frequencies when the technique is used of encrypting only a zero frequency coefficient of a DCT of luminance values for a block of pixel locations. Although embodiments have been shown wherein data in the colour image signal is either encrypted or not encrypted at all, it should be understood that alternatively the data that has been described as not encrypted at all may be encrypted with some simple encryption technique that does not require significant computational effort for encryption and/or decryption (or in any case less computational effort than the encryption and/or decryption of the data that has been described as encrypted). For example, the data that is described as unencrypted may be EXORed with the output of some simple random generator, while the data that is described as encrypted is computed as some "one way" function or EXORed with the output of a more complicated random generator. Thus some additional protection may be realized. However, the most efficient protection is realized if the data that is described as unencrypted is indeed not encrypted at all.

In another embodiment any DCT coefficient of luminance or colour component values is encrypted and decrypted but only for selected blocks. This embodiment differs from the earlier embodiments by not limiting the encrypting to luminance coefficients but encrypting colour components also. This increases the amount of computational effort needed for encryption and decryption however, provides more security. The blocks may be selected as described for luminance (for example every Nth block, where N is an integer greater than one, N=2 or 3 for example) or some more random selection may be made.

Claims

CLAIMS:

1. A method of forming a partially encrypted digital image signal from a source colour image signal, the method comprising

- computing first digital data by at least partially encrypting data that represents a luminance component of the source colour image signal; - inserting the first digital data in the partially encrypted digital image signal;

- inserting unencrypted second digital data, or second digital data that is encrypted less strongly than the first digital data into the partially encrypted image signal, the unencrypted or less strongly encrypted second digital data indicating all colour properties of pixel values from the source colour image signal that have luminance according to the luminance component.

2. A method according to Claim 1, comprising transmitting the partially encrypted image signal via a remote distribution medium.

3. A method according to Claim 1, comprising storing the partially encrypted image signal in a distributable storage medium.

4. A method according to Claim 1, comprising copying the source colour image signal to the partially encrypted digital image signal, except for a replacement of the at least part of the data that represents a luminance component in the source colour image signal by the encrypted first digital data.

5. A method according to Claim 1, the method comprising

- grouping pixel locations of the source colour image into a matrix of blocks, - computing for each block a zero frequency coefficient of a Discrete Cosine transform of luminance components for the pixel locations in the block and non-zero frequency coefficients of the Discrete Cosine Transform;

- encrypting data derived from the zero frequency coefficients of at least part of the blocks to form the first digital data; - including the encrypted zero frequency coefficients and unencrypted data derived from the non-zero frequency coefficients in the partially encrypted image signal.

6. A method according to Claim 5, the method comprising selecting a subset of the blocks for the encryption of data derived from the zero frequency coefficients, and inserting all luminance information of the blocks outside the subset into the partially encrypted digital image signal without encryption, or with weaker encryption than the encrypted blocks.

7. A method according to Claim 6, the method comprising adding information to the at least partially encrypted signal that indicates which of the blocks belong to the subset and which not.

8. A method according to Claim 6, comprising - computing a measure of deviation between the source colour signal and a result of decoding the partially encrypted signal without decryption, and

- adapting a relative frequency of blocks in the subset relative to an overall block frequency until at least a threshold value of said measure is reached.

9. A method according to Claim 1, the method comprising encrypting only part of the bits of numbers that represents a luminance component of respective blocks of the source colour image signal.

10. A method of generating an image signal from a partially encrypted digital image signal, the method comprising

- selectively decrypting a first part of the partially encrypted image digital signal that represents a luminance component;

- extracting a remaining part of the partially encrypted image digital signal without decryption or with weaker decryption than said first part, the remaining part including all colour properties of pixel values from the source colour image that have luminance according to the luminance component;

- forming a luminance component and a colour component of the image signal from the decrypted first part and the remaining part digital data respectively.

11. A method according to Claim 10, the method comprising

- forming an intermediate signal and copying the partially encrypted image signal into the intermediate signal except for said first part, and inserting a decrypted version of the first part in instead of the part in the intermediate signal; - using the intermediate signal to form the image signal without further decryption.

12. A method according to Claim 10, the method comprising selectively decrypting only a zero frequency component of Digital Cosine Transforms of the luminance component for at least part of a plurality of blocks into which the image is subdivided.

13. A method according to Claim 12, the method comprising selecting a subset of the blocks for said decryption and decoding all luminance information of the blocks outside the subset without decryption or with weaker decryption than for the blocks of the subset.

14. A method according to Claim 12, the method comprising

- reading information from the partially encrypted signal that indicates which of the blocks belong to the subset and which not, and

- decrypting selected ones of the blocks under control of said information and using the decrypted selected ones of the blocks to generate the image, using other blocks than the selected ones without decryption or with the weaker decryption to generate the image.

15. An image encryption apparatus, the apparatus comprising

- an input (10) for receiving a source colour image signal;

- an output (19) for outputting a partially encrypted digital image signal; - an encoding circuit (12, 14a,b, 16, 18) coupled between the input (10) and the output (19), comprising an encryption circuit (16) arranged to encrypt first digital data that is derived from a luminance component of the source colour image signal and to insert the encrypted first digital data in the partially encrypted digital image signal, the encoding circuit (12, 14a,b, 16, 18) being arranged to insert unencrypted second digital data or second data that is less strongly encrypted than the first digital data into the partially encrypted image signal, the unencrypted or less strongly second digital data indicating all colour properties of pixel values from the source colour image that have luminance according to the luminance component.

16. An image encryption apparatus according to Claim 15, comprising a bypass element (14a,b, 18) for copying the source colour image signal to the partially encrypted digital image signal, except for the first digital data, the encryption circuit (16) being arranged to substitute the encrypted first digital data in the place of the first digital data in the source image signal.

17. An image encryption apparatus according to Claim 15, wherein the encryption circuit (16) is arranged to derive said first digital data from a zero frequency coefficient of a Discrete Cosine Transforms of the luminance component of a subset of a set of blocks into which an image is divided, non-zero frequency coefficients of the Discrete Cosine Transform of the luminance of the blocks being inserted into the partially encrypted image signal without encryption or with weaker encryption than the blocks in the subset.

18. An image encryption apparatus according to Claim 17, comprising a feedback loop (60, 62) arranged to compute a measure of deviation between the source colour signal and a result of decoding the partially encrypted signal without decryption, and to adapt a relative frequency of blocks in the subset relative to an overall block frequency until at least a threshold value of said measure is reached.

19. An image decryption apparatus, the apparatus comprising

- an input (20) for receiving a partially encrypted digital image signal,

- an output for an output image signal;

- a decoder circuit (22, 24, 26, 27a,b) comprising a decryption circuit (24) arranged to decrypt first digital image data from a part of the partially decrypted digital image signal that represents a luminance component and to control a luminance part of the output image signal dependent on the decrypted data, the decoder circuit (22, 24, 26, 27a,b) being arranged to control all colour properties of the output image signal under control of undecrypted second digital data, or second digital data that is decrypted with weaker decryption than the first digital data, from the partially encrypted digital image signal.

20. An image decryption apparatus according to Claim 19, wherein the apparatus comprises a bypass element (27a,b) coupled between the input and the output and arranged to copy the partially encrypted image signal from the input to the output image signal except for the first digital data, the decryption circuit (26) being arranged to insert the decrypted first digital data in the place of the original first digital data among the remainder of the partially encrypted signal in the output image signal.

21. An image decryption apparatus according to Claim 19, wherein the first image data includes only a zero frequency component of Digital Cosine Transforms of the luminance component for a subset of a plurality of blocks into which the image is subdivided.

22. An image decryption apparatus according to Claim 21, comprising a control circuit (50) arranged to read information from the partially encrypted signal that indicates which of the blocks belong to the subset and which not, and to control selection between insertion of the decrypted version and copying of an original version of said part dependent on said information.

23. An image decryption apparatus according to Claim 19, wherein the decryption circuit (12, 16) arranged to decrypt only part of the bits of numbers that represents a luminance component of respective blocks of the source colour image signal.

24. A computer program product, comprising instructions which, when executed by a programmable apparatus will make the programmable apparatus perform the method of claim 1.

25. A computer program product, comprising instructions which, when executed by a programmable apparatus will make the programmable apparatus perform the method of claim 10.

26. A partially encrypted digital image signal comprising

- an at least partially encrypted luminance component of a source colour image;

- unencrypted digital data, or digital data that is encrypted less strongly than the encrypted luminance component indicating all colour properties of pixel values from the source colour image that have luminance according to the luminance component.

27. A method of forming a partially encrypted digital image signal from a source image signal, the method comprising

- grouping pixel locations of the source image into a matrix of blocks, - computing for each block a zero frequency coefficient of a Discrete Cosine transform of image components for the pixel locations in the block and non-zero frequency coefficients of the Discrete Cosine Transform;

- encrypting data derived from the zero frequency coefficients of a subset of the blocks to form the first digital data;

- including the encrypted zero frequency coefficients and unencrypted data derived from the non-zero frequency coefficients in the partially encrypted image signal;

- inserting all information of the blocks outside the subset into the partially encrypted digital image signal without encryption.

28. A method according to Claim 27, the method comprising adding information to the at least partially encrypted signal that indicates which of the blocks belong to the subset and which not.

29. A method according to Claim 27, comprising

- computing a measure of deviation between the source signal and a result of decoding the partially encrypted signal without decryption, and

30. A method of generating an image signal from a partially encrypted digital image signal, the method comprising

- selecting a subset of the blocks for decryption;

- selectively decrypting a part of the partially encrypted image digital signal that represents an image component;

- decrypting all image information of the blocks outside the subset without decryption.

31. A method according to Claim 30, the method comprising

- decrypting selected ones of the blocks under control of said information and using the decrypted selected ones of the blocks to generate the image, using other blocks than the selected ones without decryption to generate the image.

32. An image encryption apparatus, the apparatus comprising

- an input for receiving a source image signal;

- an output for outputting a partially encrypted digital image signal;

- an encryption circuit coupled between the input and the output and arranged to encrypt first digital data that is derived from a component of the source image signal from a zero frequency coefficient of a Discrete Cosine Transforms of a subset of blocks into which an image is divided, and to insert the encrypted first digital data in the partially encrypted digital image signal;

- a further circuit coupled between the input and the output and arranged to insert unencrypted second digital data into the partially encrypted image signal, the unencrypted further digital data indicating all properties zero frequency coefficients of the Discrete Cosine Transform of the blocks outside the subset.

33. An image encryption apparatus according to Claim 32, comprising a feedback loop arranged to compute a measure of deviation between the source signal and a result of decoding the partially encrypted signal without decryption, and to adapt a relative frequency of blocks in the subset relative to an overall block frequency until at least a threshold value of said measure is reached.

34. An image decryption apparatus, the apparatus comprising

- an input for receiving a partially encrypted digital image signal,

- an output for an output image signal;

- a decryption circuit arranged to decrypt first digital image data from a part of the partially decrypted digital image signal that represents a subset of blocks into which the image is divided and to control an output image signal dependent on the decrypted data;

- a further circuit arranged to control all properties of the output image signal under control of undecrypted second digital data from the blocks outside the subset in the partially encrypted digital image signal.

35. A computer program product, comprising instructions which, when executed by a programmable apparatus will make the programmable apparatus perform the method of claim 27.

36. A computer program product, comprising instructions which, when executed by a programmable apparatus will make the programmable apparatus perform the method of claim 30.