Pulse or digital communications – Bandwidth reduction or expansion – Television or motion video signal
Reexamination Certificate
2000-04-28
2002-09-17
Le, Vu (Department: 2613)
Pulse or digital communications
Bandwidth reduction or expansion
Television or motion video signal
C375S240080
Reexamination Certificate
active
06452970
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The invention relates to a method and a device for processing a digitized image.
2. Description of the Related Art
In the digital transmission of an image, particularly following image compression—that is to say, the reduction of image. data that are to be transmitted—the channel over which the transmission is accomplished is a bottleneck. The image data are transmitted over this channel from an (image) encoder to an (image) decoder. The bandwidth of this channel is mostly predetermined and constant, that the compression is correspondingly tailored to this bandwidth in the coder. According to the possibilities opened up by the standardized block-based image coding methods, for instance MPEG 4 or H.263 (see International Telecommunication Union, Telecommunications Sector of ITU, Draft ITU-T Recommendation H.263, Videocoding for low bitrate communication, May 2, 1996), the coder can be preset, or respectively, adapted, such that the number of images transmitted per unit of time is guaranteed, in exchange for which a loss of image quality is accepted.
One example is the videotelephone, which displays images that are transmitted over an ISDN B-channel with a transmission rate of 64 Kbit/s and that comprise a corresponding image quality: The images always produce an abrupt sequence having low resolution, and the images are displayed relatively small.
Block-based image coding methods (e.g. in accordance with MPEG or H.263 standards) are known from J. De Lameillieure, R. Schafer MPEG-2-Bildcodierng für das ‘digitale Fernsehen’, Fernseh-und Kino-Technik.
In image processing, the image quality is specified for a complete image that is to be transmitted. The image quality is modified depending on the available data rate, so that the bandwidth is exploited and the entire image can be transmitted. The disadvantage here is that a loss of quality must be accepted for the entire transmitted image.
SUMMARY OF THE INVENTION
It is the object of an invention to create a method and an arrangement for image processing whereby the above described disadvantage is avoided.
The object is achieved by a method for image processing, in which a digitized image is divided into two regions, whereby a first region satisfies a predetermined criterion, and a second region does not satisfy the predetermined criterion; in which the first region is processed with a higher image quality than the second region; in which the image quality is achieved in that the first region is quantized using a first quantization value, and the second region is quantized using a second quantization value, the first quantization value being less than the second quantization value; in which, when the criterion is satisfied, an optimally large contiguous region comprising several blocks that has the predetermined color (skin color) is guaranteed in that one of the following possibilities is realized: when a block does not satisfy the criterion, and this block is largely surrounded by additional blocks that do satisfy the criterion, this block is also coded with the first quantization value if the result of the comparison operation lies slightly over the predetermined threshold value; and when a block does not satisfy the criterion, and this block is situated amid blocks that do satisfy the criterion, then this block is also coded with the first quantization value.
The invention also provides an arrangement for image processing with a processor unit which is laid out such that a digitized image is divided into two regions, whereby a first region satisfies a predetermined criterion, and a second region does not satisfy the predetermined criterion; the first region is processed with a higher image quality than the second region; the image quality is achieved in that the first region is quantized using a first quantization value, and the second region is quantized using a second quantization value, the first quantization value being less than the second quantization value; in which, when the criterion is satisfied, an optimally large contiguous region comprising several blocks which has the predetermined color (skin color) is guaranteed in that one of the following possibilities is realized: when a block does not satisfy the criterion, and this block is largely surrounded by additional blocks that do satisfy the criterion, this block is also coded with the first quantization value if the result of the comparison operation lies slightly over the predetermined threshold value; and when a block does not satisfy the criterion, and this block is situated amid blocks that do satisfy the criterion, then this block is also coded with the first quantization value.
A method for image processing is designed by which a digitized image is divided into two regions: a first region which satisfies a predetermined criterion, and a second region which does not satisfy the predetermined criterion. The first region is then processed with a higher image quality.
An advantage of this is that a better image quality is achieved for the first region, which satisfies the predetermined criterion. The invention thus makes it possible to electronically process—that is, to transmit or compress—the first region as a part of the total image with a higher quality.
A development of the invention consists in dividing the image into several regions, whereby a number of first regions that satisfy the criterion are processed with a higher image quality than a number of second regions that do not satisfy the criterion.
This makes it possible to subdivide numerous regions within an image into several first and several second regions, and thus also to assign a higher image quality to noncontiguous regions of the same type as the first regions in the image.
A possible development of the invention is realized in that the criterion is satisfied when the first region of the image has a predetermined color. This can be a color similar to the color of human skin, for example.
A further development is realized in that the image is processed using a block-based image coding method. Examples of block-based image coding methods are methods that are defined in accordance with the MPEG or H.263 standards.
In a block-based coding method, for a block of the image of a predetermined size the color of the block is determined in the form of second color values, preferably by averaging the image points of this block. A comparison operation of the second color values to the first color values is performed. If a result of the comparison operation is less than a predetermined threshold value, the criterion is satisfied; that is, the block has a color that is at least similar to human skin. Otherwise (that is, the result of the comparison operation is not less than the predetermined threshold value), the criterion is not satisfied for this block.
A further development consists in carrying out the method iteratively for each block of the image.
The predetermined size of the block preferably amounts to 8×8 image points or 16×16 image points.
The comparison operation can be defined in different ways. Three possibilities are given below (see equations (1) to (3)):
|
x
y
−h
y
|+|x
Cr
−h
Cr
|—|x
Cb
−h
Cb
|<S
(1),
|
x
y
−h
y
|
2
+|x
Cr
−h
Cr
|
2
+|x
Cb
−h
Cb
|
2
<S
(2),
k
i
·|D
1
|+k
2
·|D
2
|+k
3
·|D
3
|<S,
(3),
where
x
y
references a luminance value (=brightness) of the first color values;
x
Cr
references a first chrominance value (=color tone) of the first color values;
x
Cb
references a second chrominance value (=color saturation) of the first color values;
h
y
references a luminance value (=brightness) of the second-color values;
h
Cr
references a first chrominance value (=color tone) of the second color values;
h
Cb
references a second chrominance value (=col
Le Vu
Siemens Aktiengesellschaft
LandOfFree
Method and device for processing a digitized image does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and device for processing a digitized image, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and device for processing a digitized image will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2870526