Image analysis – Image compression or coding – Transform coding
Reexamination Certificate
1994-11-02
2003-04-08
Tran, Phuoc (Department: 2621)
Image analysis
Image compression or coding
Transform coding
C382S239000
Reexamination Certificate
active
06546145
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates to an image compressing apparatus, such as an image compressing apparatus having an orthogonal transformation function, and to the method of image compression used therein.
2. Description of the Related Art
The memory capacity necessary for storing a full color image (hereinafter referred to as an “image”) such as a photograph in a memory is calculated by multiplying the number of pixels by the number of tone bits. This makes necessary an enormous memory capacity in order to store such a high-quality color image. For this reason, a variety of methods of compressing the amount of information have been proposed. For example, attempts have been made to reduce the required memory capacity by first compressing the image information and then storing the compressed information in memory.
FIG. 1
is a block diagram showing the coding method (see “International Standardization for Color Photographic Coding”, Hiroshi Yasuda, The Journal of the Institute of Image Electronics Engineers of Japan, Vol. 18, No. 6, pp. 398-407, 1989) of a baseline system proposed by the JPEG (Joint Photographic Experts Group) as a method of achieving international standardization of color still-picture coding.
As shown in
FIG. 1
, pixel data entered from an input terminal
1
is cut into 8×8 pixel blocks in a block forming circuit
2
, the data is subjected to a cosine transformation by a discrete cosine transformation (hereinafter referred to as “DCT”) circuit
17
, and the transformation coefficients obtained by the transformation are supplied to a quantization (hereinafter referred to as “Q”) unit
40
. In accordance with quantization-step information supplied by a quantization table
41
, the Q unit
40
subjects the transformation coefficients to linear quantization. Of the quantized transformation coefficients, a DC coefficient is applied to a predictive coding circuit [hereinafter referred to as a “DPCM (differential pulse-coded modulation) circuit”]
42
, which obtains the differential (a prediction error) between this DC coefficient and the DC component of the preceding block. The difference is applied to a one-dimensional Huffman coding circuit
43
.
FIG. 2
is a block diagram showing the details of the DPCM
42
. The quantized DC coefficient from the Q unit
40
is applied to a delay circuit
53
and a subtracter
54
. The delay circuit
53
applies a delay equivalent to the time needed for the discrete cosine transformation circuit to operate on one block, namely 8×8 pixels. Accordingly, the delay circuit
53
supplies the subtracter
54
with the DC coefficient of the preceding block. As a result, the subtracter
54
outputs the differential (prediction error) between the DC coefficient of the current block and that of the preceding block. In this predictive coding, the value of the preceding block is used as the prediction value, and therefore the predicting unit is constituted by the delay circuit, as set forth above.
In accordance with a DC Huffman code table
44
, the one-dimensional Huffman coding circuit
43
applies variable-length coding to the prediction error signal supplied by the DPCM
42
and supplies a multiplexer
51
with the data, i.e., a DC Huffman code, that has been variable-length coded.
An AC coefficient (a coefficient other than the DC coefficient) quantized by the Q unit
40
is zigzag-scanned in order from coefficients of lower order, as shown in
FIG. 3
, by means of a scan converting circuit
45
, and the output of the scan converting circuit
45
is applied to a significant-coefficient detector circuit
46
. The latter determines whether the quantized AC coefficient is “0” or not. If the AC coefficient is “0”, a count-up signal is supplied to a run-length counter
47
, thereby incrementing the counter. If the coefficient is “1”, however, a reset signal is applied to the run-length counter
47
to reset the counter, and the coefficient is split into a group number SSSS and annexed bits, as shown in
FIG. 4
, by a grouping circuit
48
. The group number SSSS is supplied to a two-dimensional Huffman coding circuit
49
, and the annexed bits are supplied to the multiplexer
51
. The run-length counter
47
counts a “0” run length and supplies the two-dimensional Huffman coding circuit
49
with the number NNNN of consecutive “0” between significant coefficients other than “0”. In accordance with the AC Huffman code table
50
, the Huffman coding circuit
49
applies variable-length coding to the “0” run length NNNN and the significant-coefficient group number SSSS of significant coefficients and supplies the multiplexer
51
with the data, i.e., an AC Huffman code, that has been variable-length coded.
The multiplexer
51
multiplexes the DC Huffman code, AC Huffman code and annexed bits of one block (8×8 input pixels) and outputs the multiplexed data, namely compressed image data, from its output terminal
52
. Accordingly, the compressed data outputted by the output terminal
52
is stored in a memory, and at read-out the data is expanded by a reverse operation, thereby making it possible to reduce memory capacity.
However, the example of the prior art described above has a disadvantage. For example, consider application of the prior art to an image output unit. In general, an image output unit often is connected to an image input unit such as a host computer or image scanner and operates as part of a system. In such case, various images, such as computer graphics produced by the host computer or images inputted from the image scanner are sent to the image output unit.
The prior art described above is such that a deterioration in image quality is suppressed in an image of the kind in which the transformation coefficients concentrate in the low region of the orthogonal transformation, as in an image obtained by digitizing an image such as a photograph by an image scanner. However, in artificially created images such as computer graphics, font images and images resulting from computer-aided design, images that have been compressed and then restored by expansion experience a major deterioration in quality.
Though means has been proposed in U.S. patent application Ser. No. 07/738,562 by the same assignee as this case for changing over the quantization conditions adaptively depending upon the level of transformation coefficients, this proposal also possesses a drawback. Specifically, since a sampled image is dealt with as an input source, the arrangement is such that the sensed portion of the image is distinguished as being an edge portion or a flat portion of the image on the basis of the transformation coefficients. The input source, i.e., the sampled image, which has entered from a device such as an image scanner is outputted as an image in which edge portions are weakened, no matter how significant the edge portions are in the original, owing to the characteristics of the MTF (modulation transmission function) of the image scanner. Consequently, image quality will not be affected that much even if the high-frequency components within a block are quantized coarsely to a small degree. However, an artificially produced image often contains strong AC power in high-frequency components which do not occur at the edge portions of an image inputted by a device such as an image scanner, namely an image having ordinary low to intermediate resolution. When the conventional coarse quantization is applied to these images, deleterious effects become conspicuous, such as the interruption of artificially produced fine lines and the occurrence of noise such as ringing in flat portions in the vicinity of fine lines. In addition, a simple method which involves little time loss has not been proposed with regard to the requirements for changing over the quantization conditions based upon the transformation coefficients.
The prior art described has the following drawbacks as well.
A half-tone image obtained by inputting an original such as a photograph using a device such as an image scann
Miyake Nobutaka
Nakayama Tadayoshi
Fitzpatrick ,Cella, Harper & Scinto
Tran Phuoc
LandOfFree
Image compression using selection of quantization method does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Image compression using selection of quantization method, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Image compression using selection of quantization method will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3097803