JPS63169878A - Vector encoding system for image data - Google Patents

Abstract

PURPOSE:To reduce the time required for vector encoding vector data by comparing a vector of a block to be processed with a vector of a block whose encoding is finished to select the succeeding processing thereby encoding the vector. CONSTITUTION:A difference between a vector Xi of a block to be processed inputted by a difference discrimination device 1 and a just preceding vector Xi-1 whose processing is finished is obtained and an index of the vector Xx-i is outputted as a code when the difference is within a prescribed threshold value. A standard data vector made in advance and an index corresponding to the vector of each standard data are stored in a data set section 6, which is connected to a matching processing circuit 5. When the matching processing circuit 5 is started by a start signal outputted from the matching start circuit 4, the standard data vector and its index are read sequentially from the data set section 6 to apply matching processing to the vector Xi inputted from a shift register R1.

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field] The present invention relates to a vector encoding method for image data used in image information transmission or image processing.

[Prior art]

In order to reduce the amount of data to be transmitted in image transmission, the image data to be transmitted is divided into a large number of blocks of, for example, 8 x 8 pixels, and vectors and data for each pixel in these blocks are created in advance. Performs matching processing with the vector of each standard data in a data set consisting of multiple standard data, selects the standard data having the most similar vector to the vector of the processed block, and attaches it to the selected standard data. Vector quantization is performed to select the code that is selected, and the code is transmitted. On the receiving side, standard data is selected from the data set prepared in advance using the transmitted code, and the block pattern is determined using the vector of this standard data. is being restored.

Figure 2 is a diagram conceptually showing a method for encoding such an image. For example, the original image A to be processed is a digital image in which image data such as brightness and color are quantized for each pixel. For example, the original image is divided into n blocks S 1.32 + ..., S, 1 for each 8×8 pixel.
The image data of each pixel included in each of these blocks x1. Vector X consisting of x2゜..., x64
In order to encode positive values, m standard data Y are vectorized by setting the quantization value of each pixel of 8×8 in advance.
1゜Y2 , ..., Y, are prepared, and the block to be processed S
For the vector X1 of the standard data 9Y1., and the vectors Yi, Y2°, .
Y2. . . , Yl, a matching process is performed for each block S1 . S2. ...+Sn.

In this way, when the standard data having the vector Y with the smallest amount of error with respect to the vector can get.

However, conventionally, when performing the above-described matching process on the image data of each block, all standard data vectors Y1. Y2 +..., Y power is referenced, so it is possible to obtain the same result as the matching process performed earlier because the brightness or color is almost constant, such as the background of an image, and there is almost no difference in the image data of each block. Even when the difference is large, matching processing is performed by referring to all the standard data, so there is a problem that processing time is wasted.

〔the purpose〕

The present invention focuses on the above-mentioned problems and aims to shorten the time required for processing vector encoding of image data.

〔composition〕

FIG. 1 is a diagram for explaining the present invention in detail, and FIG.
The image is divided into a large number of blocks S as shown in (B) of the same figure, and the vector encoding process described above is completed up to the vector The index corresponding to the standard data is coded a, b.
, c... shows the state obtained by closing.

Such an image may include many parts, such as the background of the image, where the brightness or color is almost constant and there is almost no difference in the vectors of each block. Vector xi of processing flock Si
The code obtained by vector encoding is the code e of the block for which processing has been completed, for example, the immediately preceding block 5i-1.
Or another adjacent block S j-1* Sj
* Since there is a high possibility that the code will match a code such as 37+1, vector encoding can be achieved by comparing the vector of the block to be processed with the vector of the block that has already been processed and using the code that has already been obtained. processing time is reduced.

FIG. 3 is a diagram showing a circuit according to an embodiment of the present invention, in which vector encoding is performed by comparing the block to be processed with the vector of the block whose encoding processing has been completed immediately before the block to be processed.

Assuming that vectors X1 to xi-+ are currently encoded, shift registers R1 and R2 contain vectors X and -
, l X, -, are respectively latched, and vector X1-1 is transferred from shift register R1 to shift register R2
is output, and furthermore, a vector X1-2 is output from the shift register R2 to one input terminal of the difference determiner 1.

When the vector Xi of the block to be processed is input to the input terminal T, a clock pulse is applied to the shift registers R1 and R2, the vector Xi is latched by the shift register R1, and is output to one input terminal of the difference determiner 1. Vector Xi- which was output from shift register R1 at the same time
, are latched by the shift register R2 and output to the other input terminal of the difference determiner 1.

This difference determiner 1 calculates the difference between the input vector Xi of the processed block and the immediately preceding vector X1-1 that has already been processed, and if the difference is within a certain value, the gate A pulse is output, this gate 2 is made conductive, and the index of the vector X1-1, which was obtained in the previous process and has already been latched in the latch circuit 3, is output as a sign.

If the difference determined by the difference determiner 1 exceeds the dark value, a pulse is output to the matching starting circuit 4, and a starting signal is sent from the matching starting circuit 4 to the matching processing circuit 5.
is output to.

This matching processing circuit 5 is connected to a data set section 6 that stores standard data vectors prepared in advance and indexes corresponding to each standard data vector, and receives a starting signal output from the matching starting circuit 4. When the matching processing circuit 5 is activated by The index of the vector The index will be output.

FIG. 4 shows a circuit according to another embodiment of the present invention, which calculates the difference between the vector of the block to be processed and the vectors of a plurality of blocks for which encoding processing has been completed, respectively.
It outputs the index of the processed block whose difference is the smallest within a certain darkness value.

When the vector Xi of the block to be processed is inputted from the input terminal T to one input terminal of the difference detection circuit 11, the data buffer 16 in which the processed vector is stored in advance is processed according to the address output from the address control circuit 17. The accessed and processed vectors Xk are sequentially output from the data buffer 16 and sequentially input to the other input terminal of the difference detection circuit 11. Furthermore, the indexes I of the processed vectors Xk that are sequentially outputted from the data buffer 16 are simultaneously and sequentially outputted to the launch circuit 21 from the index buffer 18 in which the indexes are stored in advance.

When the processed vectors X are sequentially input to the difference detection circuit 11, the differences between the vector Xi of the processed block and each processed vector
2 and the difference determination circuit 15.

This difference determination circuit 15 determines whether the input difference is within a threshold value, and if the difference is within the dark value, it outputs a strobe signal to the latch circuit 12 and output from the difference detection circuit 11. The latched difference is latched in the launch circuit 12, and the latched difference is output to the latch circuit 13 and inputted to one input terminal of the comparator 14.

At this time, the latch circuit 13 has already latched the smallest difference found so far and input it to the other input terminal of the comparator 14, so this minimum difference and the above The comparator 14 compares the difference in magnitude with the difference input from the latch circuit 12.

If the difference input from the latch circuit 12 is smaller, the comparator 14 outputs a strobe signal to the latch circuit 13 and the latch circuit 21, and the difference output from the latch circuit 12 is latched to the latch circuit 13. At the same time, the index of the processed vector when calculating the difference outputted from the latch circuit 12, that is, the index Ik outputted from the index buffer 18, is latched by the latch circuit 21 and sent to a subsequent processing circuit (not shown). is output to.

The vector
The index obtained as a result of encoding the vector is stored in the index buffer 18.

Note that when all the differences determined by the difference determination circuit 15 exceed the dark value, a start signal is output from the difference determination circuit 15 to the matching processing circuit 19, and the matching processing circuit 19 is started by this startup signal. Then, the standard data vectors and their indices are sequentially read out from the data set unit 20 similar to that explained with reference to FIG. 3, and matching processing is performed on the vector Xi input from the input terminal T. The obtained index is stored in the index buffer 18 and outputted via the latch circuit 21.

As mentioned above, the encoded vector and index are stored in the data buffer 16 and index buffer 18, and are stored in the same manner as above when encoding the vector of the next block to be processed. It is referenced and processed.

In this embodiment, the vectors and indexes read from the data buffer 16 and the index buffer 18 are selected by the access address output by the address control circuit 17, so by controlling this address control circuit 17, for example, the target This can be done by selecting vectors of blocks that are frequently used in other neighboring areas of the block or inner edges of already processed vectors.

〔effect〕

As explained above, according to the present invention, the vector of the block to be processed is compared with the vector of the block for which encoding has been completed, and the subsequent processing is selected and the vector is encoded. , it is possible to shorten the time required for vector encoding processing of image data.

[Brief explanation of the drawing]

Fig. 1 is a diagram explaining the present invention in detail, Fig. 2 is a diagram conceptually showing the encoding process to which the present invention is applied, Fig. 3 is a diagram showing an embodiment of the invention, Fig. 4 FIG. 3 is a diagram showing another embodiment of the present invention.

Claims (5)

[Claims] (1) Divide the image into a large number of blocks each containing a certain number of pixels, and compare a vector consisting of data of each pixel included in these blocks with a plurality of preset standard data vectors for the block, In a vector encoding method for image data in which a code corresponding to standard data having a vector most similar to the vector of the processed block as a result of the comparison is encoded as the code of the processed block, the vector of the processed block and the A vector encoding method for image data, which is characterized in that the vector is compared with the vector of a block for which processing has been completed, and subsequent processing is selected and the vector is encoded. (2) When the difference between the vector of the block to be processed and the vector of the immediately preceding block for which the encoding process has been completed is less than a certain threshold, the code of the block for which the encoding process has been completed is applied. The vector encoding method for image data according to claim 1, characterized in that the code is a code of a processing block. (3) A code of a block whose encoding process has been completed such that the difference between the vector of the processed block and the vector of a certain number of blocks whose encoding process has been completed is the minimum below a certain threshold 2. The vector encoding method for image data according to claim 1, wherein: is the code of the block to be processed. (4) The encoding process is completed such that the difference between the vector of the block to be processed and the vector of the block whose encoding process has been completed in the vicinity of this block on the image is the minimum below a certain threshold. 4. The vector encoding method for image data according to claim 3, wherein the code of the block being processed is used as the code of the block to be processed. (5) The code of the block in which the difference between the vector of the block to be processed and the vector of a certain number of blocks in the order of frequency of use among blocks whose encoding processing has been completed is the smallest and is less than a certain threshold value is determined as the code of the block to be processed. 4. A vector encoding method for image data according to claim 3, characterized in that the image data is encoded as follows.