Image analysis – Color image processing – Compression of color images
Reexamination Certificate
1997-09-15
2004-01-06
Johnson, Timothy M. (Department: 2625)
Image analysis
Color image processing
Compression of color images
C375S240030, C382S165000, C382S239000, C382S251000
Reexamination Certificate
active
06674897
ABSTRACT:
BACKGROUND OF THE INVENTION
This invention relates to a picture data compressing device which may be advantageously employed in an equipment handling compressed picture data, such as a digital Video Tape Recorder (VTR), television telephone system, teleconference system or a transmitter for a television station, and a red data detection device employed in such picture data compressing apparatus.
There has been known a digital VTR for digital recording/reproduction of picture information data. The recording system of the digital VTR is arranged as shown for example in
FIG. 13
, in which audio signals are fed via an input terminal
50
to an A/D converter
51
and picture signals are fed via an input terminal
53
to an A/D converter
54
.
The A/D converter
51
digitizes the audio signals to generate audio data which is supplied to an audio recording processing circuit
52
which then processes the audio data into a form suitable for recording and transmits the processed audio data to an error correction encoding circuit
61
.
The A/D converter
54
digitizes the picture signals by, for example, frame-based sampling, to generate component data, that is luminance data (Y data) and two color difference data (R-Y data and B-Y data), and transmits the component data to a blocking circuit
56
within a compression encoding circuit
55
.
The blocking circuit
56
divides each of the Y-data, R-Y data and the B-Y data into blocks each consisting of a matrix of 8 vertically arrayed pixels and 8 horizontally arrayed pixels (DCT blocks), these blocks being routed to a shuffling circuit
57
.
The shuffling circuit
57
generates a single macro-block from 8 DCT blocks, namely 6 Y data blocks, 1 R-Y data block and 1 B-Y data block. These 8 blocks are located at the same location on a picture. The shuffling circuit
57
effects pre-set shuffling on the macro-block basis, generates a single unit from five macro-blocks and outputs data on the unit basis. The unit picture data is supplied to a discrete cosine transform (DCT) circuit
58
.
The DCT circuit
58
transforms the picture data of each macro-block making up the unit with data on the time scale into that on the frequency scale, and routes the transform coefficients to a quantization circuit
59
.
The quantization circuit
59
selects such quantization coefficients which will give a fixed data length of the unit picture data outputted from a variable length encoding circuit
60
as later explained, and re-quantizes the unit-based picture data with the aid of these quantization coefficients. The unit-based picture data thus re-quantized is supplied to the variable length encoding circuit
60
.
The variable length encoding circuit
60
processes the unit-based data into fixed-length data so that the data quantity of the picture data is not more than a pre-set quantity, and routes the resulting fixed-length data to the error correction coding combining circuit
61
.
The error correction coding combining circuit
61
combines the unit-based picture data having the fixed length with the audio data supplied from the audio data recording processing circuit
52
to generate recording data. In addition, the error correction encoding combining circuit
61
appends so-called parity data for error correction to the recording data and routes the resulting signals to a recording modulation circuit
62
.
The recording modulation circuit
62
modulates the recording data from the combining operation in a pre-set manner and routes the resulting modulated signals to a recording head
63
. The recording head
63
records the picture data obliquely on a video tape, not shown.
By digitally recording the picture data (and audio data) in this manner, recording/ reproduction may be achieved without deteriorating the picture quality by preventing adverse effects of the random noise or the like.
However, the above-mentioned digital VTR has a drawback that, since the picture data are divided into plural macro-blocks and processed with DCT or re-quantization or the like on the macro-block basis, there arises the risk of occurrence of so-called block distortion, in which the noise is produced at a junction between macro-blocks on the reproduced picture. Above all, the block distortion of a macro-block containing a large quantity of red-hued picture data, tends to be visually outstanding, and thus has been desired to be improved.
SUMMARY OF THE INVENTION
In view of the above-described problem of the prior art, it is an object of the present invention to provide a picture data compressing apparatus in which the block distortion containing a large quantity of the red-hued picture data is diminished to contribute to improvement in the picture quality.
It is another object of the present invention to provide a red data detection device in which the number of bits required in determining whether or not the picture data is red-hued may be reduced for simplifying the hardware.
The present invention provides a picture compressing device including blocking means for dividing picture data into plural blocks each consisting of a pre-set number of blocks and outputting the resulting blocked picture data, and transform encoding means for transform encoding picture data from the blocking means on the block basis and outputting resulting transform coefficients. The picture compressing device also includes red block detecting means for detecting the block transmitted thereto from the blocking means as being a red block if the block has more than a pre-set number of picture data of red-hued pixels, and quantization means for quantizing transform coefficients from the transform coding means. The picture compressing device additionally includes control means for variably controlling the quantization steps of the quantization means if the block is detected as being a red-hued block by the red block detection means.
With the picture data compressing device according to the present invention, the controlling means variably controls the quantization step of the quantization means so that the quantization step will become finer if the block is detected as being a red-hued block by the red block detection means.
With the picture data compressing device according to the present invention, the controlling means variably controls the quantization step of the quantization means quantizing red data so that the quantization step will become finer for a macro-block consisting of a luminance data block and two color data blocks.
With the picture data compressing device according to the present invention, controlling means performs activity-based classing for a pre-set number of macro-blocks as a processing unit, and adaptively controls the quantization steps of the quantization means according to classes so that such class is selected in which the quantization step of the quantization means quantizing red-hued data is finest if the block from the blocking means is found to be the red-hued block by the red block detection means.
The present invention also provides a red data detecting, device including upper bit extracting means for extracting upper three bits of red-hued data supplied as 8-bit bi-level data and outputting the extracted upper three bits, threshold outputting means for outputting, as threshold data, upper three bits of 8-bit bi-level data that is closest to a reference value employed in detecting red data and that may be divided out by a power of 2, where an exponent is a natural number, and comparator means for comparing red-hued data from the upper bit extracting means with threshold data from the threshold data outputting means. The comparator means outputs high-level red detection data indicating that the picture data is the red data or the low level red detection data indicating that the picture data is not red data when the red data is larger or smaller than the threshold data, respectively.
The present invention also provides a red data detecting device including upper most bit extracting means for extracting upper most bit of blue-hued data supplied as 8-bit bi-level data and
Fischl Jason
Izumi Nobuaki
Kiyama Yuka
Sugisaki Kiminori
Yanagihara Naofumi
Frommer William S.
Frommer & Lawrence & Haug LLP
Johnson Timothy M.
Sony Corporation
LandOfFree
Picture data compression device and red data detection device does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Picture data compression device and red data detection device, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Picture data compression device and red data detection device will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3204345