JP2862233B2 - Information transmission system - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to an information transmission system, and more particularly to a system for transmitting information by switching the information density to be transmitted for each information group obtained by dividing the information to be transmitted by a predetermined amount. It is about. <Conventional technology> When transmitting information, the theme is always how to reduce the amount of information to be transmitted so that the original information can be faithfully reproduced. For this reason, various transmission methods have been proposed. Have been. For the above-mentioned theme, there is an adaptive variable density sampling method in which the sampling density, that is, the information density to be transmitted is appropriately changed. As an example of this method, the time axis conversion band compression method (TAT; Time Axis Transf
orm) will be briefly described below. FIG. 5 is a diagram for explaining the basic concept of TAT. The original signal is divided every predetermined period (predetermined amount of information) as indicated by a dotted line, and it is determined whether the information included in each divided group is coarse or dense. For the group determined to be dense, all of the data obtained by sampling the original signal is transmitted as transmission data, and for the block determined to be coarse, only a part of all data is regarded as transmission data, and the others are interposed. It is not transmitted as reference data. According to the above-described concept, the number of data transmitted per unit time is reduced, and the band of the transmission signal can be compressed. The data transmitted in this manner is used on the receiving side to form data corresponding to the thinned data. That is, interpolation data that approximates the thinning data is calculated using the transmitted data. Since the interpolation data corresponds to a coarse portion of the information, the interpolation data is very similar to the thinned data. Therefore, the fidelity of the restored signal to the original signal is hardly changed as compared with the case where all data is transmitted, and the transmission band can be greatly compressed. That is, the amount of information to be transmitted can be reduced. On the other hand, for each block, whether to transmit all sampling data or a part of the data is determined by examining the details of the original signal, and this determination information is also transmitted as transmission mode information in some form simultaneously. . Now, let us consider applying the above-described concept to transmission of image information. Since the image information has a two-dimensional spread and has correlation in both horizontal and vertical directions,
If not only the sampling interval in the horizontal direction but also the sampling interval in the vertical direction is made variable, more effective transmission becomes possible. This concept is hereinafter referred to as a two-dimensional TAT, which will be briefly described below. The concept of the two-dimensional TAT has already been disclosed in Japanese Patent Application No. 60-148112 and the like. FIG. 6 is a diagram showing a data transmission pattern in a two-dimensional TAT. In 2D TAT, one screen is mxn
The pixel data is divided into pixel blocks each composed of a pixel and the density of transmission data is changed for each pixel block. In FIG. 6, two types of transmission modes are shown, assuming that the pixel block has 4 × 4 pixels. In the figure, ○ indicates a transmission pixel, and × indicates a thinned pixel.
E indicates a pattern for transmitting all pixel data as shown, and C indicates a pattern for transmitting only a part of all pixel data. Hereinafter, transmission modes based on these transmission patterns are referred to as E mode and C mode, respectively. As is clear from the figure, it can be understood that the C mode performs transmission at an information density of 1/4 of the E mode. With respect to the decimated pixels transmitted in the C mode in the pixel block, the interpolated pixel data is formed on the receiving side by using the pixel data adjacent thereto in the transmitted pixel data, and the original screen is restored. Hereinafter, a configuration for realizing such transmission by two-dimensional TAT will be described. FIG. 7 is a diagram showing a schematic configuration example on the transmission side of a transmission system using two-dimensional TAT. Seventh
In the example shown in the figure, a digital transmission system is described as an example. The input video signal is converted to an analog / digital converter (A /
D) All pixels are sampled in 1 to generate all pixel data. When all the pixel data is supplied to the thinning circuit 2, after performing band limitation, thinning corresponding to the C mode pattern in FIG. 6 is performed to obtain C mode data. The C-mode data is supplied to the interpolation circuit 3, and the interpolation pixel data corresponding to the thinned pixel is calculated. The interpolated pixel data is supplied to the mode discriminating circuit 4 together with all the pixel data output from the A / D1, and it is determined whether each pixel block is transmitted in the C mode or the E mode. The mode discriminating circuit 4 calculates the difference between the pixel data output from the A / D1 and the interpolated pixel data, calculates the sum of the differences (hereinafter referred to as block distortion) for each pixel block, and calculates this for one field. Store it in memory. Until the next field data is input, the distribution of block distortion of all pixel blocks is obtained.
Here, the ratio between the number of pixel blocks transmitted in the C mode and the number of pixel blocks transmitted in the E mode and the number of pixel blocks transmitted in the E mode must always be constant. For example, C
If the number of pixel blocks transmitted in the mode is set to 2/3 of the total number of pixel blocks and the number of pixel blocks transmitted in the E mode is set to 1/3 of the total, the total number of data to be transmitted (compression ratio) is (2/3 × 1/4 + 1) /
3 × 1 =) 1/2. Therefore, a distortion threshold value for determining the degree of the block distortion at which the C mode and the E mode are to be distributed is obtained from the distribution of the block distortion of all the pixel blocks. Then, the stored block distortion is sequentially read out at the timing when the video signal of the next field is input, and the transmission mode is determined by comparing with the distortion threshold. When the read block distortion matches the distortion threshold, the transmission mode is set so that the ratio of the pixel block transmitted in the C mode and the pixel block transmitted in the E mode at a predetermined ratio as described above matches. It is determined. The mode discrimination signal obtained as described above is supplied to the switch 7, and pixel data is read out from the buffer 52 for E mode data and the buffer 6 for C mode data. The output data of the switch 7 is output to the transmission line as transmission data. The mode discrimination signal is also output to the transmission line as mode information via the buffer 9. FIG. 8 is a diagram showing a schematic configuration example on the receiving side of a two-dimensional TAT transmission system. The digital video signal that has been subjected to the above-described processing and that is input via a transmission path is output from a C-mode interpolation circuit 11.
And the interpolation data corresponding to the thinned pixel data in the C mode is calculated. On the other hand, the transmitted mode information controls the switch 12,
When the mode information indicates the E mode, the terminal is connected to the E side, and when the mode information indicates the C mode, the terminal is connected to the C side. Thereby, all the pixel data including the E mode data, the C mode data and the interpolated pixel data are stored in the frame memory 13. For example, all pixel data is read out from the frame memory 13 in an order conforming to a television signal, and is output through a digital / analog converter (D / A) 14. As described above, in a two-dimensional TAT transmission system, image information can be transmitted very effectively. <Problems to be solved by the invention> In the transmission system of the two-dimensional TAT described above, the bandwidth is reduced by half.
Although it is possible to reduce the amount of information, the amount of information per unit time is about 5 times as large as that of the current NTSC television signal when transmitting a so-called high-definition television signal that has been attracting attention in recent years. Increase. Therefore, it is impossible to transmit the data through a transmission path such as a magnetic recording / reproducing system unless the data amount is further reduced. The present invention provides an information transmission system that can further reduce transmission information by the conventional variable-density sampling method and hardly impairs the fidelity of transmission information to original information in the background described above. With the goal. <Means for Solving the Problems> Under the above-described object, the present invention provides an information transmission system for transmitting image information of each information group obtained by dividing image information for one screen to be transmitted for each predetermined pixel. And a first transmission unit that outputs only representative data obtained by encoding data of one or more representative pixels in each of the information groups as image information of each of the information groups, In each information group, the additional data and the representative data obtained by quantizing a difference value between the data obtained from the representative data and the data other than the one or more pixels in a predetermined quantization step are the respective information groups. A second transmission unit that outputs the image information of the group together with the data related to the quantization step, and image information and image data restored using the image information output from the first transmission unit. Output means for selectively outputting the output of the first transmission means or the second transmission means for each of the information groups in accordance with a difference between each of the information groups with respect to null image information. And transmitting the image information output from the transmission means and the second transmission means in a single screen, and reducing the number of quantization bits of the additional data to the number of quantization bits of the representative data. An information transmission system, which is a feature of the present invention, is provided. <Operation> According to the configuration as described above, the number of transmission bits of additional data can be further reduced in addition to the transmission information amount reduction effect obtained by variable density sampling, and it is necessary to perform nonlinear quantization. Therefore, the maximum value of the quantization error can be reduced, so that the fidelity of the information to be transmitted to the original signal can be obtained with almost no deterioration, and information transmission with extremely high transmission efficiency can be performed. <Example> An example of the present invention will be described below. FIG. 1 is a diagram showing a schematic configuration of a transmission side of a transmission system as one embodiment of the present invention. In FIG. 1, the same components as those in FIG. In the transmission system shown in FIG. 1, when transmission is performed in the E mode, a normal quantization bit number is given to C-mode data (basic pixel data) that is not band-limited, and pixels other than the basic pixel data are assigned. The (additional pixel) data is to be transmitted with a smaller number of quantization bits. Here, the additional pixel data is transmitted as data obtained by quantizing a difference value from the basic pixel data by an adaptively determined quantization step. The E-mode data and C are stored in the field memories 25 and 26, respectively.
The mode data is stored, and the E mode data is further guided to the switch 7 via the prediction difference encoder 27. FIG. 2 is a diagram showing a specific configuration of the encoder 27 in FIG. The E-mode data input from the field memory 25 to the encoder 27 is supplied to a switch 45. Here, the address number of each data is defined. FIG. 3 is a diagram schematically showing address numbers of data stored in the field memories 25 and 26 in FIG. I in the figure is 4 ×
The small block number 4 is shown as shown in FIG. 3 (a). For the E mode data in the small block having the small block number i, the address number of the upper left pixel, that is, the basic pixel is E (i, 1). , The upper right, lower left, and lower right pixels have address numbers of E (i, 2), E (i, 3), and E (i, 4), respectively. The address number of the C-mode data having the small block number i is C
(I). The switch 45 stores the basic pixel data of the address number E (i, 1) in the basic pixel buffer memory 38 and the address number E (i, 1).
2) Switching is performed so that the additional pixel data of E (i, 3) and E (i, 4) can be stored in the additional pixel buffer memories 42,43,44, respectively. In the quantization range determination circuit 47, E within one small block
The difference between the basic pixel data and the additional pixel data is calculated using the four pixel data of the address numbers (i, 1) to E (i, 4), respectively, and the peak value or the average value is used. If the peak value or the average value is large, the quantization step of the DPCM encoder 40 is determined so as to increase the quantization step. Now, the E-mode data input to the encoder 27 is 8 bits, and the difference additional pixel data to be output from the encoder 27 is 4 bits. The difference between the two E-mode data is 9-bit data. When four consecutive bits are extracted from the 9 bits, there are six patterns, and up to six types of quantization steps can be selected. A better image can be transmitted if there are many types of quantization steps that can be set here.However, the number of bits of range data indicating the quantization step increases, so the number of quantization steps that can be set depends on the screen to be transmitted. Should be determined. The switches 46 are sequentially stored in the memory 25 as E (i, 2), E (i, 3), E
Additional pixel data with the address number (i, 4) is sequentially DPC
It is supplied to the M encoder 40. The DPCM encoder 40 quantizes the difference between the basic pixel data and the additional pixel data in each small block in the quantization step determined by the determination circuit 47, and supplies the quantized data to the I terminal of the switch 41. The switch 41 selectively outputs, as E-mode data, basic pixel data and DPCM data in which a quantization step is adaptively determined corresponding to additional pixel data. The quantization bit range determination circuit 47
The output range bit data is transmitted separately. FIG. 4 is a flowchart for explaining the above-mentioned operation in FIGS. 1 and 2. j is a small block number starting from j = 0, and it is first determined from the mode information of the small block whether the small block transmitted in the C mode is a small block transmitted in the E mode. If it is determined that the mode is the E mode, the switch 45 in FIG. 2 is connected to the terminal A, and then the address number E (j,
The basic pixel data of 1) is accessed and written into the basic pixel buffer memory (BM) 38. Next, after the switch 45 is connected to the terminal B, the address number E (j, 2) is written to the additional pixel BM42 by accessing the additional pixel data. Switch 45
Is connected to the terminal C, the additional pixel data of the address number E (j, 3) is accessed to the additional pixel BM43, and the switch 45 is connected to the terminal D, and the additional pixel data of the address number E (j, 4) is Access and write to the additional pixels BM44 respectively. After the quantization range is determined by the quantization range determination circuit 47, the switch 41 is connected to the H side and the switch 7 is connected to the E side to output the basic pixel data of the address number E (j, 1).
Next, the switch 46 is connected to the E terminal, and the address number E (j,
The DPCM data △ (j, 2) corresponding to the additional pixel data of 2) is calculated. Then, the switch 41 is connected to the I side, and △ (i, 2) is output from the switch 7. Next, the switch 46 is connected to the F terminal to calculate and output the DPCM data △ (j, 3) corresponding to the additional pixel data of the address number E (j, 3). Finally, the switch 46 is connected to the G terminal. Address number E (j,
The DPCM data △ (j, 4) corresponding to the additional pixel data of 4) is calculated and output, j is set to j + 1, and the process proceeds to the next small block. If it is determined that the small block to be transmitted is in the C mode, the C mode field memory C (j) is read out, the switch 7 is connected to the C side, and the process proceeds to the next small block. These timings are set so that one data is always output from the switch 7 at the same interval.
The read timing is controlled. If the number of quantization steps is four and the range data is two bits for one small block, the data compression ratio is (1/3 × 11/16) + (2/3 × 1/4) = 19/48, which enabled data compression of about 40%. Moreover, the image quality hardly changes as compared with the case where the conventional variable density sampling is performed. In the above-described embodiment, as a method of predictive difference encoding, a method of nonlinearly quantizing data of a difference between basic pixel data and each additional pixel data in a small block of (4 × 4) is used. The encoding method is not limited to this. For example, in the (4 × 4) small block, data on the difference between the average value of two basic pixel data located on the right and left is used for the upper right additional pixel data, and the additional pixel data on the lower left is located on the upper and lower sides. It is also possible to use data of the difference between the average value of the two basic pixel data and the additional pixel data at the lower right using data of the difference from the average value of the basic pixel data located on all sides. is there. Further, in the E mode, it is also possible to transmit C mode data and data of a difference between the C mode data and each pixel data. Further, in the above-described embodiment, each pixel block is transmitted in either the C mode or the E mode. Blocks can also be provided. Details of this are disclosed in Japanese Patent Application No. 60-230510 filed by the present applicant. Further, in the above-described embodiment, the information to be transmitted is a video signal. However, the present invention can be applied to the case of transmitting other information having some correlation on the time axis, for example, an audio signal. <Effect of the Invention> As described above, according to the present invention, it is possible to further reduce the amount of information to be transmitted without deteriorating the fidelity to the original signal as compared with the case where information is transmitted by the conventional variable density sampling method. As a result, it has become possible to obtain an information transmission system capable of transmitting a large amount of information per unit time over a relatively narrow band transmission line. Further, according to the present invention, each of the information groups transmitted by any of the transmission means is transmitted including the representative data, and the data by the first transmission means and the second transmission means are mixed and transmitted. Even if a transmission error occurs in some of the information blocks, it is possible to restore a minimum of each information group and to make the transmission error inconspicuous.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a diagram showing a schematic configuration on a transmission side of a transmission system as one embodiment of the present invention, and FIG. 2 is a diagram showing a specific configuration example of an encoder in FIG. FIG. 3 is a diagram schematically showing address numbers of data stored in the field memory shown in FIG. 1, and FIG. 4 is a diagram showing an operation of sending and receiving data of each memory shown in FIG. 1 and FIG. FIG. 5 is a diagram for explaining conventional information transmission by variable density sampling, FIG. 6 is a diagram showing a data transmission pattern in a two-dimensional TAT, and FIG. 7 is a transmission of a transmission system using a two-dimensional TAT. FIG. 8 is a diagram showing a schematic configuration example of a receiving side of a two-dimensional TAT transmission system. In the figure, 2 is a thinning circuit, 4 is a mode discriminating circuit, 7 is a switch, 38 is a basic pixel buffer memory, 42, 43, and 44 are additional pixel buffer memories, 40 is a DPCM encoder, and 47 is a quantization range determining circuit.

   ────────────────────────────────────────────────── ─── Continuation of front page    (72) Inventor Soichi Kashida               770 Shimonoge, Takatsu-ku, Kawasaki-shi Canon               Tamagawa Works Co., Ltd. (72) Inventor Shin Shimokoriyama               3-30-2 Shimomaruko, Ota-ku, Tokyo               Inside Canon Inc. (72) Inventor Naoto Abe               3-30-2 Shimomaruko, Ota-ku, Tokyo               Inside Canon Inc. (72) Inventor Nobuyoshi Yamashita               3-30-2 Shimomaruko, Ota-ku, Tokyo               Inside Canon Inc.                (56) References JP-A-52-38813 (JP, A)                 JP-A-59-79651 (JP, A)                 "Technical Report of the Institute of Television Engineers of Japan (Vo               l. 8, No. 2) "1984. 4 p. 47               −54                 "Digital Signal Processing of Images"               Business Newspaper p. 146-147

Claims (1)

(57) [Claims] An information transmission system for transmitting image information of each information group obtained by dividing image information for one screen to be transmitted for each predetermined pixel, wherein one or a plurality of representative pixels in each of the information groups is provided. A first transmission unit for outputting only representative data obtained by encoding the data of the information group as image information of each of the information groups; and in each of the information groups, data obtained from the representative data; A second transmission unit that outputs the additional data and the representative data obtained by quantizing a difference value with data other than the pixel in a predetermined quantization step together with the data related to the quantization step as image information of each information group. And the first transmission according to the difference between the image information restored using the image information output from the first transmission means and the original image information for each information group. Means for selectively outputting the output of the means or the second transmission means for each information group, wherein the image information output from the first transmission means and the second transmission means are mixed in one screen An information transmission system, wherein the number of quantization bits of the additional data is smaller than the number of quantization bits of the representative data.