WO2010018494A1

WO2010018494A1 - Image compression

Info

Publication number: WO2010018494A1
Application number: PCT/IB2009/053431
Authority: WO
Inventors: Shuo Li; Abraham Karel Riemens; Tomas Henriksson; Pieter Van Der Wolf
Original assignee: Nxp B.V.
Priority date: 2008-08-11
Filing date: 2009-08-06
Publication date: 2010-02-18

Abstract

An image compression unit for compressing image data (ID) having a plurality of pixels, each with one or more color samples, into a compressed format (CF) is provided. The image compression unit comprises a color lookup table generation unit (311) for generating a color lookup table having one or more colors and for transmitting the color lookup table in the compressed format (CF). The image compression unit further comprises a pixel mode determination unit (312) for determining for at least one pixel of the image data (ID) a pixel mode and for transmitting for the at least one pixel a pixel mode identifier for identifying the pixel mode in the compressed format (CF). A first pixel mode corresponds to the case that the at least one pixel matches one of the one or more colors in the color lookup table and a second pixel mode corresponds to the case that the at least one pixel does not match one of the one or more colors in the color lookup table. The image compression unit further comprises a color lookup table encoding unit (313) for encoding pixels whose pixel mode is the first pixel mode and a predictive encoding unit (314) for predictively encoding pixels whose pixel mode is the second pixel mode.

Description

IMAGE COMPRESSION

FIELD OF THE INVENTION

The present invention relates to an image compression unit, an image decompression unit, an image compression method, and an image decompression method. The invention further relates to an embedded image processing system.

BACKGROUND OF THE INVENTION

A. K. Riemens, et al., "Transparent Embedded Compression in Systems-on- Chip", Proc. of IEEE Workshop on Signal Processing Systems Design and Implementation, Banff, AB, Cananda, Oct. 2006 describes an image compression unit integrated in an embedded image processing system. The image compression unit uses a predictive encoding, wherein the input signal is split into two parts: an MSB part, which is losslessly compressed using a delta pulse code modulation (DPCM) and variable length coding (VLC), and an LSB part, which is lossy.

This image compression algorithm works well on natural image contents, however, when executed on graphics type image contents, e.g. on graphics overlays, artefacts may occur leading to a reduced decompressed image quality.

SUMMARY OF THE INVENTION

It is an object of the present invention to provide an image compression unit, an image decompression unit, an image compression method, and an image decompression method, which work well on both natural and graphics type image contents. It is a further object of the present invention to provide a corresponding image processing system.

This object is solved by an image compression unit according to claim 1, an image decompression unit according to claim 14, an image compression method according to claim 13, an image decompression method according to claim 16, and an embedded image processing system according to claim 17.

Therefore, an image compression unit for compressing image data having a plurality of pixels, each with one or more color samples, into a compressed format, is provided. The image compression unit comprises a color lookup table generation unit for generating a color lookup table having one or more colors and for transmitting the color lookup table in the compressed format, and a pixel mode determination unit for determining for at least one pixel of the image data a pixel mode and for transmitting for the at least one pixel a pixel mode identifier for identifying the pixel mode in the compressed format. A first pixel mode corresponds to the case that the at least one pixel matches one of the one or more colors in the color lookup table. A second pixel mode corresponds to the case that the at least one pixel does not match one of the one or more colors in the color lookup table. The image compression unit further comprises a color lookup table encoding unit for encoding pixels whose pixel mode is the first pixel mode, and a predictive encoding unit for predictively encoding pixels whose pixel mode is the second pixel mode. The invention relates to the observation that in many graphics type image contents merely a few colors, e.g. the colors of a writing or a pictogram in the foreground of the graphic, occur often in the data set. By storing these colors in a color lookup table, the pixels that match one of these colors can be encoded efficiently by the color lookup table encoding. By treating these pixels separately, the remaining pixels, e.g. the pixels of the background of the graphic, are more likely to have similar color values or to exhibit only smooth changes in color. It is therefore more likely that these pixels can be encoded efficiently using predictive encoding.

According to an embodiment, the image data that is compressed into the compressed format is stored in a memory in said compressed format. According to an embodiment, the size of the color lookup table is substantially smaller than the maximum number of colors that can be represented by the signal range of the image data.

According to an embodiment, the color lookup table generation unit is adapted for generating the color lookup table based on the most occurring colors of the image data. By doing so, a preferably large number of pixels can be encoded efficiently by the color lookup table encoding unit.

According to a further embodiment, the pixel mode determination unit is adapted for transmitting the pixel mode identifier as at least one bit. A first state of the at least one bit indicates the first pixel mode and a second state of the at least one bit indicates the second pixel mode. Hence, the pixel mode identifier can be transmitted as a single bit as this provides the highest compression efficiency.

According to a further embodiment, the color lookup table encoding unit is adapted for encoding pixels whose pixel mode is the first pixel mode as an index in the color lookup table. This provides a very efficient encoding method as the number of bits required for representing an index in a color lookup table is typically much smaller than the number of bits required to represent the corresponding colors directly.

According to a further embodiment, the predictive encoding unit is adapted for encoding pixels whose pixel mode is the second pixel mode with a prediction delay. The prediction delay is dynamically adjusted to skip pixels whose pixel mode is the first pixel mode.

According to a further embodiment, the predictive encoding unit is adapted for encoding pixels whose pixel mode is the first pixel mode by a delta pulse code mo dulation scheme .

According to a further embodiment, the predictive encoding unit is adapted for splitting the color samples of pixels whose pixel mode is the first pixel mode in most significant bits and least significant bits and for encoding the most significant bits of said pixels. According to an embodiment, the length of the color lookup table can be adapted based on the colors of the image data. The length of the color lookup table is transmitted in the compressed format. By making the length of the color lookup table adaptable, the length can be chosen to be optimal for the image data to be compressed. According to a further embodiment, the image compression unit is adapted such that the encoding of pixels by the color lookup table encoding unit can be deactivated. This allows for an adaptation to both natural and graphics type image content.

According to an embodiment, the color lookup table generation unit is adapted for storing only the most significant bits of colors in the color lookup table. This reduces the required size of the color lookup table.

The invention also relates to an image compression method for compressing image data having a plurality of pixels, each with at least one color sample, into a compressed format. A color lookup table having one or more colors is generated and the color lookup table is transmitted in the compressed format by a color lookup table generation unit. For at least one pixel of the image data a pixel mode is determined and for the at least one pixel a pixel mode identifier for identifying the pixel mode is transmitted in the compressed format by a pixel mode identification unit. A first pixel mode corresponds to the case that the at least one pixel matches one of the one or more colors in the color lookup table. A second pixel mode corresponds to the case that the at least one pixel does not match one of the one or more colors in the color lookup table. Pixels whose pixel mode is the first pixel mode are encoded by a color lookup table encoding unit. Pixels whose pixel mode is the second pixel mode are predictively encoded by a predictive encoding unit. The invention also relates to an image decompression unit for decompressing image data having a plurality of pixels, each with one or more color samples, from a compressed format. The image decompression unit comprises a color lookup table reading unit for reading a color lookup table having one or more colors from the compressed format, and a pixel mode identification unit for identifying for at least one pixel of the image data a pixel mode by reading for the at least one pixel a pixel mode identifier for identifying the pixel mode from the compressed format. A first pixel mode corresponds to the case that the at least one pixel matches one of the one or more colors in the color lookup table. A second pixel mode corresponds to the case that the at least one pixel does not match one of the one or more colors in the color lookup table. The image decompression unit further comprises a color lookup table decoding unit for decoding pixels whose pixel mode is the first pixel mode, and a predictive decoding unit for predictively decoding pixels whose pixel mode is the second pixel mode.

According to an embodiment, the image data that is decompressed from the compressed format is retrieved from a memory in said compressed format.

The invention also relates to an image decompression method for decompressing image data having a plurality of pixels, each with at least one color sample, from a compressed format. A color lookup table having one or more colors is read from the compressed format by a color lookup table reading unit. For at least one pixel of the image data a pixel mode is identified by reading for the at least one pixel a pixel mode identifier for identifying the pixel mode from the compressed format, by a pixel mode identification unit. A first pixel mode corresponds to the case that the at least one pixel matches one of the one or more colors in the color lookup table. A second pixel mode corresponds to the case that the at least one pixel does not match one of the one or more colors in the color lookup table. Pixels whose pixel mode is the first pixel mode are decoded by a color lookup table decoding unit. Pixels whose pixel mode is the second pixel mode are predictively decoded by a predictive decoding unit. The invention also relates to an embedded image processing system for processing image data having a plurality of pixels, each with at least one color sample. The embedded image processing system comprises one or more image processing units for processing the image data. The embedded image processing system further comprises an image compression unit for compressing the image data to a compressed format as defined in anyone of the claims 1 to 12, and/or an image decompression unit for decompressing the image data from the compressed format as defined in anyone of the claims 14 to 15.

It shall be understood that the image compression unit of claim 1 , the image compression method of claim 13, the image decompression unit of claim 14, the image decompression method of claim 16, and the embedded image processing system of claim 17 have similar and/or identical embodiments as defined in the dependent claims.

It shall be understood that an embodiment of the invention can also be any combination of the dependent claims with the respective independent claims.

BRIEF DESCRIPTION OF THE DRAWINGS

These and other aspects of the invention will be apparent from and elucidated with reference to the embodiment(s) described hereinafter. In the following drawings: Fig. 1 shows a block diagram of an image processing system for processing image data according to a first embodiment,

Fig. 2 shows a block diagram of an image compression unit for compressing image data to a compressed format according to a second embodiment,

Fig. 3 shows a flowchart illustrating an image compression method for compressing image data to a compressed format according to the invention,

Fig. 4 shows a block diagram of an image decompression unit for decompressing image data from a compressed format according to a third embodiment, and

Fig. 5 shows a flowchart illustrating an image decompression method for decompressing image data from a compressed format according to the invention.

DETAILED DESCRIPTION OF EMBODIMENTS

Image data comprise a plurality of pixels (short for picture element) representing the rows and columns of a digital image. Depending on the color format used to represent the image, e.g. greyscale, RGB, YUV, or aRGB (RGB with alpha), the image data comprises one or more color channels. For example, an image in greyscale format comprises a single luminance channel. An image in RGB format comprises three color channels, namely a red (R) color channel, a green (G) color channel, and a blue (B) color channel. An image in aRGB format comprises, in addition to the three color channels of an RGB image, a fourth alpha channel that can be used to provide a transparency value, e.g. for supporting blending effects. Each pixel of an image comprises one or more color samples, each representing a value of a respective color channel. For example, a pixel in a greyscale image comprises a single luminance sample representing a value of the luminance channel. A pixel in an RGB image comprises three color samples, each representing a value of the red (R), green (G), and blue (B) color channel, respectively. A pixel in an aRGB image comprises, in addition to the color samples of an RGB image, a fourth alpha sample representing, for example, a transparency value. Fig. 1 shows a block diagram of an image processing system for processing image data ID according to a first embodiment. The image processing system can be an embedded system implemented as a System-on-Chip (SoC). The image processing system comprises a number of image processing units 100 ... 1Ox (only two are shown by way of example), an image compression unit 310, an image decompression unit 320, and a memory controller 400 interconnected via a bus 200. The embedded image processing system can further comprise a memory 600, for example, an off- chip DRAM or SDRAM. Alternatively, the memory 600 can be implemented on-chip. The memory controller 400 controls the transfer of data between one or more image processing units 100 ... 10x and the memory 600 via a memory interface 500. Because bandwidth in particular to off-chip memory is a scarce resource in complex SoCs, it is an important goal to be able to reduce the amount of data to be transferred to and from the memory. In order to alleviate this bandwidth bottleneck, the image compression unit 310 compresses image data ID "on-the-fly" into a compressed format CF when written to the memory 600. Correspondingly, the image decompression unit 320 decompresses image data ID "on-the-fly" from the compressed format CF when read from the memory 600. The image data ID that is compressed by the image compression unit 310 can comprise complete images. Alternatively, images can be compressed in smaller chunks of data. For example, an image can be compressed row by row, or in rectangular blocks of e.g. 8x8 pixels. The image data ID in these cases correspond to the individual rows of the image or to the rectangular pixel blocks. Furthermore, for images with more than one color channel, the different color channels can also be compressed separately. For example, an image in RGB format can also be compressed as three separate greyscale images where the luminance channel of the first greyscale image corresponds to the red (R) color channel of the RGB image, the luminance channel of the second greyscale image corresponds to the green (G) color channel of the RGB image and the luminance channel of the third greyscale image corresponds to the blue (B) color channel of the RGB image. The image compression unit 310 and the image decompression unit 320 can be separate units. Alternatively, they can be integrated into a single image compression/decompression unit 300.

Fig. 2 shows a block diagram of an image compression unit 310 according to a second embodiment. The image compression unit 310 according to the second embodiment can be used in the image processing system according to the first embodiment. The image compression unit 310 is adapted for compressing image data ID to a compressed format CF in accordance with the invention. The image compression unit 310 comprises a color lookup table generation unit 311 for generating a color lookup table having one or more colors and for transmitting the color lookup table in the compressed format CF. The image compression unit 310 further comprises a pixel mode determination unit 312 for determining for at least one pixel of the image data ID a pixel mode and for transmitting a pixel mode identifier for the at least one pixel for identifying the pixel mode in the compressed format CF. Here, a first pixel mode corresponds to the case that the at least one pixel matches one of the one or more colors in the color lookup table and a second pixel mode corresponds to the case that the at least one pixel does not match one of the one or more colors in the color lookup table. The image compression unit 310 further comprises a color lookup table encoding unit 313 for encoding pixels whose pixel mode is the first pixel mode. A predictive encoding unit 314 serves to predictively encode pixels whose pixel mode is the second pixel mode.

In the following an image compression method for compressing image data ID to a compressed format CF according to the invention will be described with reference to a flow-chart shown in Fig. 3. In step S311, the color lookup table generation unit 311 of the image compression unit 310 generates from the image data ID a color lookup table having one or more colors and transmits the color lookup table in the compressed format CF, e.g. in a header or another suitable section of the compressed format CF. In this embodiment, the color lookup table generation unit 311 is adapted to generate the color lookup table from the most occurring colors of the image data ID, wherein the image data ID may comprise a complete image. Alternatively, in case that the image data ID is compressed in smaller chunks of data, the image data ID may correspond to the individual rows of the image or to rectangular pixel blocks. A color histogram analysis can be performed for finding those colors with the highest occurrence rate in the image data ID. These colors can be stored in the color lookup table. The color lookup table may contain substantially fewer colors than the available colors in the image data ID, e.g., only 1, 2, 4 or 8 colors.

In step S312, the pixel mode determination unit 312 of the image compression unit 310 determines for at least one pixel of the image data ID a pixel mode and transmits for the at least one pixel a pixel mode identifier for identifying the pixel mode in the compressed format CF. A first pixel mode corresponds to the case that the at least one pixel matches one of the one or more colors in the color lookup table. A second pixel mode corresponds to the case that the at least one pixel does not match one of the one or more colors in the color lookup table. The pixel mode determination unit 312 can be adapted for transmitting the pixel mode identifier as at least one bit. A first state of the at least one bit may indicate the first pixel mode and a second state of the at least one bit indicates the second pixel mode. The pixel mode identifier can be transmitted as a single bit. In step S313, the color lookup table encoding unit 313 encodes the pixels matching one of the colors in the color lookup table, i.e. the pixels that were determined to belong to the first pixel mode in step S312. Here, the color lookup table encoding unit 313 is adapted for encoding said pixels as an index in the color lookup table. In step S314, the predictive encoding unit 314 predictively encodes the pixels not matching one of the colors in the color lookup table, i.e. the pixels that were determined to belong to the second pixel mode in step S312.

In this embodiment, the predictive encoding unit 314 is adapted for predictively encoding the pixels whose pixel mode is the second pixel mode by a delta pulse code modulation (DPCM), i.e. instead of encoding each pixel directly, a difference between a pixel and a previous pixel is encoded. The calculation of pixel differences DP(k) can be described by the following equation:

DP (k) = P(k) - P(k - d) , where P(k) is the pixel to be encoded, k is the position of said pixel in the image data ID, P(k - d) is a previous pixel that is used as predictor, and k - d is the position of said previous pixel in the image data ID. The distance between P(k) and predictor P(k - d) depends on the choice of the prediction delay d.

The prediction delay d can be predefined to have a fixed value that is independent of the image data ID to be compressed. However, the prediction delay d can also be adapted to the image data ID to be compressed. In this case, the prediction delay d is transmitted in the compressed format CF, e.g. in a header or another suitable section of the compressed format CF, as an additional parameter necessary for the decompression of the compressed format CF. According to the invention, when moving to the next pixel to be encoded, the predictor location k - d is only updated when the updated predictor location corresponds to a pixel not matching one of the one or more colors in the color lookup table, i.e. when the updated predictor location corresponds to a pixel whose pixel mode is the second pixel mode. In other words, the prediction delay d is dynamically adjusted during encoding to skip pixels whose pixel mode is the first pixel mode.

The pixel differences DP(k) can be encoded by a suitable means, e.g. by a form of variable length coding (VLC) such as Rice coding. VLC coding parameters necessary for the decompression of the compressed format CF are transmitted in the compressed format CF, e.g. in a header or another suitable section of the compressed format CF.

In the above described embodiment, image data ID is compressed by combining the encoding of pixels matching one of the one or more colors in the color lookup table, i.e. pixels whose pixel mode is the first pixel mode, with the predictive encoding of pixels not matching one of the one or more colors in the color lookup table, i.e. pixels whose pixel mode is the second pixel mode. In another embodiment, the encoding of pixels by the color lookup table encoding unit 313 can be deactivated and it can be determined dynamically if the combination of color lookup table encoding and predictive encoding is suited for the image data ID to be compressed, or if it is more efficient to predictively encode all pixels of the image data ID. In this case, a compression mode identifier for identifying a first compression mode, i.e. the combination of color lookup table encoding and predictive encoding, and a second compression mode, i.e. predictive encoding only, is transmitted in the compressed format CF, e.g. in a header or another suitable section of the compressed format CF, as an additional parameter necessary for the decompression of the compressed format CF. The compression mode identifier can be transmitted as at least one bit. A first state of the at least one bit may indicate the first compression mode. A second state of the at least one bit may indicate the second compression mode. The compression mode identifier can be transmitted as a single bit. In case that only predictive coding is used, the color lookup table as well as the pixel mode identifier for identifying the pixel mode of each pixel are not transmitted in the compressed format CF.

In the above described embodiment, the color lookup table contains a fixed number of colors, e.g. 1, 2, 4 or 8 colors. In another embodiment, the length of the color lookup table can be adapted. The optimal length of the color lookup table may be determined such that it is optimal for the image data ID to be compressed, e.g. based on the number of colors that occur substantially more often than all other colors in the image data ID. In this case, the length of the color lookup table is transmitted in the compressed format CF, e.g. in a header or another suitable section of the compressed format CF, as an additional parameter necessary for the decompression of the compressed format CF.

In the above described embodiment, the color lookup table encoding is combined with a predictive encoding which encodes pixels whose pixel mode is the second pixel mode by a delta pulse code modulation (DPCM) with dynamically adjustable prediction delay d. In another embodiment, the predictive encoding unit 314 can be adapted for splitting the color samples of pixels whose pixel mode is the second pixel mode in most significant bits and least significant bits and for encoding the most significant bits of said pixels by a delta pulse code modulation (DPCM) with dynamically adjustable prediction delay d. Alternatively, the above described color lookup table encoding method can be combined with other predictive encoding methods with a prediction delay d or even with non-predictive encoding methods such as a regular pulse code modulation (PCM).

In the above described embodiment, the most occurring colors in the image data ID are stored in the color lookup table. In another embodiment, only the most significant bits of said colors are stored in the color lookup table. Fig. 4 shows a block diagram of an image decompression unit 320 according to a third embodiment. The image decompression unit 320 according to the third embodiment can be used in the image processing system according to the first embodiment. The image compression unit 320 is adapted for decompressing image data ID from a compressed format CF in accordance with the invention. The image decompression unit 320 comprises a color lookup table reading unit 321 for reading a color lookup table having one or more colors from the compressed format CF. The image compression unit 320 further comprises a pixel mode identification unit 322 for identifying for at least one pixel of the image data ID a pixel mode from the compressed format CF by reading for the at least one pixel a pixel mode identifier for identifying the pixel mode from the compressed format CF. Here, a first pixel mode corresponds to the case that the at least one pixel matches one of the one or more colors in the color lookup table and a second pixel mode corresponds to the case that the at least one pixel does not match one of the one or more colors in the color lookup table. The image decompression unit 320 further comprises a color lookup table decoding unit 323 for decoding pixels whose pixel mode is the first pixel mode. A predictive decoding unit 324 serves to predictively decode pixels whose pixel mode is the second pixel mode.

In the following an image decompression method for decompressing image data ID from a compressed format CF according to the invention will be described with reference to a flow-chart shown in Fig. 5.

In step S321, the color lookup table reading unit 321 of the image decompression unit 320 reads a color lookup table having one or more colors from the compressed format CF. In step S322, the pixel mode identification unit 322 of the image decompression unit 320 identifies for at least one pixel of the image data ID a pixel mode by reading for the at least one pixel a pixel mode identifier for identifying the pixel mode from the compressed format CF. A first pixel mode corresponds to the case that the at least one pixel matches one of the one or more colors in the color lookup table. A second pixel mode corresponds to the case that the at least one pixel does not match one of the one or more colors in the color lookup table.

In step S323, the color lookup table decoding unit 323 decodes the pixels matching one of the colors in the color lookup table, i.e. the pixels that were determined to belong to the first pixel mode in step S322. In step S324, the predictive decoding unit 324 predictively decodes the pixels not matching one of the colors in the color lookup table, i.e. the pixels that were determined to belong to the second pixel mode in step S322.

Other variations of the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention, from a study of the drawings, the disclosure, and the appended claims.

In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude a plurality.

A single unit or device may fulfill the functions of several items recited in the claims. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage.

A computer program may be stored/distributed on a suitable medium, such as an optical storage medium or a solid state medium, supplied together with or as part of other hardware, but may also be distributed in other forms, such as via the Internet or other wired or wireless telecommunication systems.

Any reference signs in the claims should not be construed as limiting the scope.

Claims

CLAIMS:

1. An image compression unit for compressing image data (ID) having a plurality of pixels, each with one or more color samples, into a compressed format (CF), comprising: a color lookup table generation unit (311) for generating a color lookup table having one or more colors and for transmitting the color lookup table in the compressed format (CF), a pixel mode determination unit (312) for determining for at least one pixel of the image data (ID) a pixel mode and for transmitting for the at least one pixel a pixel mode identifier for identifying the pixel mode in the compressed format (CF), wherein a first pixel mode corresponds to the case that the at least one pixel matches one of the one or more colors in the color lookup table and a second pixel mode corresponds to the case that the at least one pixel does not match one of the one or more colors in the color lookup table, a color lookup table encoding unit (313) for encoding pixels whose pixel mode is the first pixel mode, and a predictive encoding unit (314) for predictively encoding pixels whose pixel mode is the second pixel mode.

2. The image compression unit as claimed in claim 1, wherein the image data (ID) that is compressed into the compressed format (CF) is stored in a memory (600) in said compressed format (CF).

3. The image compression unit as claimed in claim 1 or 2, wherein the size of the color lookup table is substantially smaller than the maximum number of colors that can be represented by the signal range of the image data (ID).

4. The image compression unit as claimed in anyone of the claims 1 to 3, wherein the color lookup table generation unit (311) is adapted for generating the color lookup table based on the most occurring colors of the image data (ID).

5. The image compression unit as claimed in anyone of the claims 1 to 4, wherein the pixel mode determination unit (312) is adapted for transmitting the pixel mode identifier as at least one bit, wherein a first state of the at least one bit indicates the first pixel mode and a second state of the at least one bit indicates the second pixel mode.

6. The image compression unit as claimed in anyone of the claims 1 to 5, wherein the color lookup table encoding unit (313) is adapted for encoding pixels whose pixel mode is the first pixel mode as an index in the color lookup table.

7. The image compression unit as claimed in anyone of the claims 1 to 6, wherein the predictive encoding unit (314) is adapted for encoding pixels whose pixel mode is the second pixel mode with a prediction delay, wherein the prediction delay is dynamically adjusted to skip pixels whose pixel mode is the first pixel mode.

8. The image compression unit as claimed in anyone of the claims 1 to 7, wherein the predictive encoding unit (314) is adapted for encoding pixels whose pixel mode is the first pixel mode by a delta pulse code modulation scheme.

9. The image compression unit as claimed in anyone of the claims 1 to 8, wherein the predictive encoding unit (314) is adapted for splitting the color samples of pixels whose pixel mode is the first pixel mode in most significant bits and least significant bits and for encoding the most significant bits of said pixels.

10. The image compression unit as claimed in anyone of the claims 1 to 9, wherein the length of the color lookup table can be adapted based on the colors of the image data (ID) and the length of the color lookup table is transmitted in the compressed format (CF).

11. The image compression unit as claimed in anyone of the claims 1 to 10, wherein the encoding of pixels by the color lookup table encoding unit (313) can be deactivated.

12. The image compression unit as claimed in anyone of the claims 1 to 11, wherein the color lookup table generation unit (311) is adapted for storing only the most significant bits of the colors in the color lookup table.

13. An image compression method for compressing image data (ID) having a plurality of pixels, each with at least one color sample, into a compressed format (CF), comprising the steps: generating a color lookup table having one or more colors and transmitting the color lookup table in the compressed format (CF) by a color lookup table generation unit (311), determining for at least one pixel of the image data (ID) a pixel mode and transmitting for the at least one pixel a pixel mode identifier for identifying the pixel mode in the compressed format (CF), wherein a first pixel mode corresponds to the case that the at least one pixel matches one of the one or more colors in the color lookup table and a second pixel mode corresponds to the case that the at least one pixel does not match one of the one or more colors in the color lookup table, by a pixel mode determination unit (312), encoding pixels whose pixel mode is the first pixel mode by a color lookup table encoding unit (313), and - predictively encoding pixels whose pixel mode is the second pixel mode by a predictive encoding unit (314).

14. An image decompression unit for decompressing image data (ID) having a plurality of pixels, each with one or more color samples, from a compressed format (CF), comprising: a color lookup table reading unit (321) for reading a color lookup table having one or more colors from the compressed format (CF), a pixel mode identification unit (322) for identifying for at least one pixel of the image data (ID) a pixel mode by reading for the at least one pixel a pixel mode identifier for identifying the pixel mode from the compressed format (CF), wherein a first pixel mode corresponds to the case that the at least one pixel matches one of the one or more colors in the color lookup table and a second pixel mode corresponds to the case that the at least one pixel does not match one of the one or more colors in the color lookup table, a color lookup table decoding unit (323) for decoding pixels whose pixel mode is the first pixel mode, and a predictive decoding unit (324) for predictively decoding pixels whose pixel mode is the second pixel mode.

15. The image decompression unit as claimed in claim 14, wherein the image data (ID) that is decompressed from the compressed format (CF) is retrieved from a memory (600) in said compressed format (CF).

16. An image decompression method for decompressing image data (ID) having a plurality of pixels, each with at least one color sample, from a compressed format (CF), comprising the steps: reading a color lookup table having one or more colors from the compressed format (CF) by a color lookup table reading unit (321), - identifying for at least one pixel of the image data (ID) a pixel mode by reading for the at least one pixel a pixel mode identifier for identifying the pixel mode from the compressed format (CF), wherein a first pixel mode corresponds to the case that the at least one pixel matches one of the one or more colors in the color lookup table and a second pixel mode corresponds to the case that the at least one pixel does not match one of the one or more colors in the color lookup table, by a pixel mode identification unit (322), decoding pixels whose pixel mode is the first pixel mode by a color lookup table decoding unit (323), and predictively decoding pixels whose pixel mode is the second pixel mode by a predictive decoding unit (324).

17. An embedded image processing system for processing image data (ID) having a plurality of pixels, each with at least one color sample, comprising: one or more image processing units (100 ... 1Ox) for processing the image data (ID), an image compression unit (310) for compressing the image data (ID) to a compressed format (CF) as defined in anyone of the claims 1 to 12, and/or an image decompression unit (320) for decompressing the image data (ID) from the compressed format (CF) as defined in anyone of the claims 14 to 15.