JPS621373A - Picture encoding system - Google Patents

Abstract

PURPOSE:To identify a monochromatic picture from a color picture and to prevent the increase of an unnecessary transmission time by inputting input picture information to a quantization encoding part through a means which detects the quantity of color component signal to decide whether the picture signal corresponds to the color picture or not. CONSTITUTION:Color signals R, G, and B of the input picture are inputted to a color base converting part 101 and are converted to a luminance signal Y and color difference signals I and Q. Signals Y, I, and Q are inputted to not only a cosine conversion operating part 102 to obtain cosine conversion data FY, FI, and FQ but also an ¦I¦.¦Q¦ discriminating part 105 to calculate the total sum of absolute values of signals I and Q. If this total sum exceeds a certain value, signals FY, FI, and FQ are sent to a quantization coding part 104 through a selecting part 103. If the total sum is smaller than the certain value, the signal FY is sent to the coding part 104. This encoded data is inputted to a receiving part 110 through a communication line 109 and is converted reverse to a transmitting part 100 to output signals R, G, and B.

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field of the Invention] The present invention relates to an image encoding method for still images such as facsimile, videotex, and filing, and moving images such as TV signal transmission and recording.

[Technical background of the invention and its problems] The image can be divided into a color image and a monochrome image depending on the amount of information.
; Can be broadly classified. A color image has three times the amount of image information as a monochrome image, so this transmission requires a lot of time and storage capacity.

A compression encoding method exclusively for color images is used for filing.

Figure 5 shows a color image that is cosine transformed and encoded into G, B.
The color axis conversion unit 502C is operated by two people. Here, R, G,
B is a luminance signal Y and a color difference signal 1. 'Qc transform, and these signals Y, I, and Q are subjected to a cosine transform calculation section 503 (two people perform the cosine transform, where they are divided into blocks of 16 x 16 pixels and cosine transform is performed. This transformed data is sent to a quantization encoder. 5041-, respectively.The quantized code data is passed through a transmission circuit 505, decoded by a quantization decoder 506 which is a receiving section, and inputted to an inverse cosine transform calculation section 507. .

Here, inverse cosine transformation is performed, and the inversely transformed north data is transferred to the color axis conversion unit 508 (two sets Y, I, and Q to R9G, B
C: Converted and output from the output terminal 509.

However, if a signal containing a monochrome image or a monochrome image of a color image is input using this method, this monochrome image will also be compressed and encoded as a color image. In other words, in addition to the luminance signal component Y, which is a monochrome signal component, the color signal component is also encoded.
Transmitting unnecessary information results in C;. For example, facsimiles increase transmission time and communication costs. The filing device has the disadvantage of increasing its exclusive storage capacity.

Also, when a monochrome signal is encoded using a quantization encoder (2) and the encoding bit rate is small, part of the decoding error of the decoded picture is converted into a color signal, and the edge parts etc.1; There was also the problem that images were generated.

[Object of the invention] The present invention has been made in view of the above-mentioned disadvantages C.
The purpose of this is to identify monochrome images when a monochrome image or a color image containing a monochrome image is input to a color image transmission or filing device, and to avoid unnecessary increases in transmission time and storage. The present invention provides an image encoding device that eliminates an increase in capacity.

[Effects of the invention] 1) Since a monochrome image can be processed in the transmission time C2, unnecessary processing can be eliminated. If it is a so-called transmission device, communication costs can be reduced; 7) [Embodiments of the invention] An embodiment of the present invention will be described in detail below with reference to the drawings.

FIG. 1 is a block diagram of a color image transmission device according to the present invention. This device includes a transmitting section 100 and a receiving section 110 via a communication line 109. The transmitter 100 includes a color axis converter 101 , a cosine transform calculator 102 , and a selector 10
3. The receiving section 110 is composed of a quantization decoding section 108.1 and an IQ+ determination section 105. Here, these R, G, and B are combined into a luminance signal Y and a color difference signal 1.
Q and C are converted. This conversion formula is expressed as: These Y, I, and Q are input into a cosine transformation operation 814102 (-, where they are divided into blocks of 16 x 16 pixels and cosine transformation is performed. This transformation (with two Yζ) is as follows: ty(jj): Luminance signal component Y The cosine transformed data Py("e" is calculated.Also, even if I, Q(Y), tx(L D l "Q
('e J) FI('L, Fq(u, ri) can be obtained in the same way.

On the other hand, the I and Q signals of Y, I, and Q are sent as 111/IQ1 determination section 105 signals to the selection section 103. In other words, when it exceeds a certain value, it is determined that a color signal component exists, and Y,
FT, FI, Fq representing the three signal components of I and Q,
Both signals are sent to the quantization code section 104 via the selection section 103.

If it does not exceed a certain value, it is determined that there is only a monochrome signal component, and only FY is sent to the quantization code section 104 via the selection section 103. The quantization encoder 104 performs quantization based on bit allocation C as shown in FIG. 3, for example. The encoded data from here is transmitted via the communication line 109.
The signal is input to the quantization unit 108 which is the receiving unit 110.

Here, this encoded information is decoded and input to the inverse cosine transform calculation unit 107, where it is inversely cosine transformed. This transformation (for Y) Pq(u, v) c from 'I(L
J) −'Q('l J) is obtained.

At this time, whether only the luminance signal Y was sent due to the additional information from the transmitting side or the color difference information 1. Q will also be sent to you and you will be able to find out if it is your next intention. For brightness information only, I, Q
Signal component! ('+5)=tq(!,j)=0(Y) is set, and the color axis inverse converter 106 inversely converts YIQ to RGB4 and outputs R, G, and B signals.

As described above, Embodiment C of the present invention: Accordingly, it is determined whether the image is a color image or a monochrome image, and encoding is performed accordingly. There is no need for gt2 encoding to increase transmission time. In other words, communication costs can be reduced.

[Other Embodiments of the Invention] In the above embodiments, a case was described in which it was determined whether the entire image was a color image or a monochrome image, but for example, as shown in FIG. Recording paper 40
In the case where a color photograph 41 is pasted on a part of the upper part (for two cases, it may be done as follows).

FIG. 2 is a block diagram of a color image transmission device according to another embodiment of the present invention. This device connects a transmitter 200 and a receiver 220 to a transmitter 200 and a receiver 220 via a communication line 211.
0 is the color axis conversion unit 2el, the cosine conversion calculation unit 202.
quantization section 203, M)1-MR encoding section 204,
Selection 4F1205, III-IQl Judgment #2
07 TelN is established, the receiving section 220 converts the quantization decoding section 20
6. Cosine inverse transform calculation unit 2089 Color axis inverse transform unit 209
．． It is composed of an MH and MR decoding section 210. Transmission section 20
0(-, the input image is divided into small blocks, and this block and .

The color signals are separated into R, G, and Bt by a color axis conversion unit 20.
11:, where these R, G, and B are input as a luminance signal Y and a color difference signal 1. Q and C are converted. And this Y, I
, Q are cosine-transformed by the cosine-transform calculation unit 202. The details are as shown in C above.

The cosine-transformed signal is quantized by quantization section 20352 and input to selection section 205.

On the other hand, the luminance signal Y from the color axis conversion section 201 is encoded by the MH or MR encoding section 204 for binary facsimile using the MH or MR encoding method, which is the CCITT standard encoding method, and is input to the selection section 205. be done.

These encoding methods focus on the probability distribution characteristics of the signal obtained by digitizing a black and white binary manuscript, and convert the amount of information into an equivalent C.
It was developed in order to compress the signal band while maintaining the signal strength.

Furthermore, as described above, the f2,111/IQ1 determining section 207 discriminates between color and monochrome, and this signal is input to the selecting section 205. In other words, the image is divided into 5 small blocks, and this block is coded with C codes. For color signals (or when at least one color component is detected within the block), Y, I
, Q are selected by the selection unit 205, and in the case of a monochrome signal, only the binary code output of the Y signal is selected. The encoded information signal is inversely transformed by the inverse cosine transform calculation unit 28, and the monochrome signal is inversely transformed by the MH-MR decoding unit 210.
The image data is decoded by the color axis inverse transformer 209, and then inversely transformed by the color axis inverse transformer 209 and sent to the output part. 208 and 209 are detailed above.

For example, if you input an image that includes a color photo that is the size of a normal black-and-white document, the image will be divided into two small blocks C, and the document part will be divided into MHs, which are used in binary facsimile, etc. , MR encoding was used, and by using color image encoding only for the color photograph part, it was possible to reduce the number of encoded coins by about the same amount as when the entire image was encoded in color. Also, monochrome images are separated from the beginning.Encoding error C: hue error of decoded image.

For example, it has the advantage of not causing image deterioration such as the appearance of a single color on the outline.

In the above embodiment, the three primary colors of light R, G,
In case of B, please explain about C. For example, in the case of a TV display, R, G, and B are suitable.

However, if the purpose is to record paper Cyu,
Yellow (Y) and magenta (M) are the three primary colors of subtractive color mixing.

It is suitable to use cyan (C). In this case, it can be seen that by changing the coefficients of the color axis conversion formula ((11)), it can be easily used for C two-speed.

Transformation (even if there is one, we used cosine transformation here, but other linear transformations such as adder),? - You may use transformations such as Rie.

In the above embodiment, the input terminal of the Ill/IQ1 determination unit (105 in Figure 1 and 207 in Figure 2) is a cosine transformation calculation unit (in the above embodiment, the image is divided into 16 x 16 blocks C, and each block is An example of performing a two-dimensional cosine transformation was shown in Figure 2, but the block division may be of other sizes, such as 8 x 8 or 32 x 32, or it may be 8 x 1 instead of a square matrix.
Rectangular division such as 6 may also be performed.

Furthermore, although the above embodiment has been explained using an example using linear transform encoding, other encoding methods such as PCM or vector quantization may be used.

In the above embodiment, discrimination between a color signal and a monochrome signal (: sum of absolute values of I and Q signals, maximum value MAX
(t!(j*J) I-1'Q(1,3N l may be used as iφj. In the latter case, when there is a color signal component even in C21 samples in a block, that block is color encoded. It corresponds to the way you do it.

In this way, the present invention provides a color image transmission device, a seven-eye ring device (-
; Since it detects that it is a monochrome image and applies an encoding method exclusively for seven-chrome images, it is efficient and has the characteristic that no encoding error (erroneous color information due to two errors occurs) even after decoding.

In the above embodiment, determination as to whether the image is a monochrome image or a color image is made using the I and Q components of Y, I, and Q. This device includes a transmitting section 600 and a receiving section 620 via a communication line 606. Transmission section 600
is the Fourier transform calculation unit 601, and is the child encoding unit 602
1 selection section 603. It is composed of an MH-MR encoding section 604 and a distribution measuring section 605. The receiving section 620 includes multiplexing sections 607 . Fourier inverse transform calculation unit 608. It is composed of a quantization decoding section 609 and an MH, MR decoding section 610. The input image signal is divided into three parts, and one of the input (multi-value) images t1 is Fourier-transformed by a Fourier transform calculation unit 601 which performs multi-value encoding, where j=4:〒1U. The converted NEC code is then subjected to binary encoding in the MH-MR encoding unit 604. This encoding method is based on the CCITT standard encoding method M
H (Modified Mutfman) or M R (
Modified Read). The encoded signal is sent to the selection section 603.

The other input image is input to a distribution measurement unit 605, where the distribution of the signal is measured. This will be explained in detail later. This distribution measurement unit 605 determines whether the image is a binary image or a multivalued image, and sends determination information to the selection unit 603. In C2 above, monochrome images, black images, and color images are distinguished from binary images and multivalue images based on black colors (e.g., black and gray).
The former is called a monochrome image because it is called a monochrome image.In other words, in this embodiment, for example, an image like a manuscript and an image with black gradation (a black and white photograph) are used. The encoded data selected by the selection unit 603 (binary or multilevel) is sent to the reception unit 620 via the communication line 606.

The encoded data sent to the receiving unit 620 is divided into two parts,
One to the quantization decoding section 609 and the other to the MH, MR decoding section 6
Sent to 10. The encoded data decoded by the quantization decoding section 609 is subjected to inverse Fourier transform in the inverse Fourier transform calculation section 608 and sent to the multiplexing section 607 . On the other hand, in the binary MH and MR decoding section 610, the encoded data undergoes MH or MR decoding processing and is sent to the multiplexing section 607. The inverse Fourier transform is expressed as. Multiplexing unit C: While switching and combining the multilevel image and the binary decoded image signal, the multiplexing unit C restores the same image as the input image and outputs it.

The distribution measurement unit 605 measures the frequency of appearance of each level of the input image signal. For example, it is like the nine distribution map shown in FIG. 7C. The horizontal axis represents the level of the image signal, and the vertical axis represents the frequency of appearance. When the input image is a binary image such as a document, the images are concentrated and distributed at two levels (,).

In addition, when the input image is a multivalued image, as shown in (b), there are many levels of
If two levels are concentrated, the input image is determined to be a binary image, and the determination information is sent to the selection unit 603 as described above.

Conventionally, when inputting multilevel data as shown in Figure 8 (a), the decoded output at 9 o'clock is shown in (b), and no major deterioration is observed, but (
When binary image data such as C) was input, the decoded output was an image in which C: the high frequency component 2 was significantly reduced, as shown in (d). This is a deterioration in image quality that inevitably occurs because binary images contain many high-frequency components.

However, by doing this, since a suitable encoding method is selected and encoded, the encoding efficiency for binary images is improved. Furthermore, since the high-frequency components do not deteriorate, it is possible to prevent the image quality from deteriorating.

In the above embodiment C, the image is like a monochrome image C; an image consisting of the signal of one screen (although I mentioned it above, it is like a color image, where one frame is represented by color signal components of three frames) The same can be applied to
IQ) component C is converted, and the luminance signal component C is subjected to the binary multi-value discrimination of the above embodiment, and the binary signal is binary encoded, and the multi-value signal is multi-value color signal encoded. It may also be encoded. Furthermore, the present invention can also be applied to a color image in which both binary and multivalued images are mixed (y).When all of the three color primary color signal components are binary,
A color-binary encoding method (for example, an encoding method that encodes the signal level and its duration) is used for each of the three colors. When this is not the case, that is, when at least one color signal component is multi-valued, the color multi-value encoding method (
For example, it is sufficient to use DPCM or linear transform coding method.

If the present invention is applied to a monochrome facsimile transmission device (: if applied, the binary signal part will be automatically encoded in C; high-speed binary encoding will be performed, even if part of the screen contains a C binary multi-value image). Reduces and eliminates image quality deterioration in the binary part,
Images can be transmitted in a short transmission time. Also, the present invention can be colored.
If you apply 7Axis, binary image encoding is automatically applied to simple image parts such as animations and backgrounds, so there will be no distortion of edge waveforms or 9color shifts, and the entire image will be multiplied. The image quality is higher than that when the value image is encoded, and the transmission time can be significantly reduced.

Furthermore, when the present invention is applied to the TV conference telephone system ζU, when the camera is photographing people, scenery, etc., the multi-value image is encoded using the multi-value image encoding method, and the multi-value image is encoded using the multi-value image encoding method, When copying an OHP or the like, binary encoding is performed.C: Higher speed and clearer document transmission can be achieved.

In addition, in all of the above-mentioned embodiments, 1 has been explained by taking up a transmission device such as a color facsimile that transmits signals via a transmission line, but this transmission circuit is a storage device C: if replaced, 7 Various modifications can be made without departing from the gist of the present invention, such as being able to be configured as an eyering device.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the configuration of an embodiment device according to the present invention, FIG. 2 is a block diagram showing the configuration of another embodiment device of the present invention, and FIG. 3 is a diagram showing quantization by bit allocation. FIG. 4 is a diagram showing an example of an image adapted to a device according to another embodiment of the present invention, FIG. 5 is a block diagram of a conventional device, and FIG. 6 is a block diagram showing the configuration of a device according to another embodiment of the present invention. , Figure 7 (a
) is a distribution diagram of a binary image, (b) is a distribution diagram of a multivalued image, and FIG. 101...Color axis conversion unit 102...Cosine transformation calculation unit 103...Selection unit 104...Quantization code unit 105...January/IQ+ determination unit 106...Color axis inversion Conversion unit 107... Cosine inverse transformation calculation unit 108... Quantization decoding unit 109... Communication line agent Patent attorney Noriyuki Chika (and one other person) Figure 8 Figure 4 5tyf 502 5tp3
5a(t≦〃

Claims (1)

[Claims] (1) means for converting and separating input image information into a luminance component signal and a color component signal; a means for detecting the amount of the separated color component signals to determine whether the input image signal is a color image; means for selecting and encoding both the luminance component signal and the color component signal when the image is determined to be a color image through this determination; and means for selecting and encoding only the luminance component signal when it is determined that the image is not a color image. An image encoding method characterized by: