JPS61125293A - Mosaic coding device - Google Patents

Abstract

PURPOSE:To permit the mosaic coding of original image data at high speed without any help of a man by selecting the mosaic pattern suitable for the image to be drawn automatically by calculation process to perform the mosaic coding of the image data. CONSTITUTION:For example the original image is scanned for M times by an image scanner 7 which is for 2,048 dots (bits) for one line. Quantitization is made for the analog information on each picture element of 2,048XM dots obtained by a binary circuit and the value is stored in an image memory GM8. Frame information for image compression (coding), having (n)X(m) dots [(n) and (m) are positive integers.], which has the same shape with a mosaic pattern is cut out from the image memory GM8. This information is compared with all of the individual pattern sequencially, having the same (n)X(m) dots, in a mosaic pattern dictionary stored in a RAM4. Then, the mosaic pattern which resembles best to the cut-out image data is selected. The mosaic coding of the image data is achieved by performing the above process for all of the frame data of (n)X(m) dots stored in the image memory.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a mosaic coding device, and particularly to a method for creating image data to be input to a center system in a captain system by combining mosaic patterns, for example.

With the recent remarkable progress in semiconductor technology, the capacity of memory elements has increased and the speed of processors has increased, making it possible to freely handle image data that cannot be encoded.

However, in image systems that transmit and receive image data through telephone lines, such as the Captain System that is currently in practical use, the amount of data transmitted is large, so the usage time of the telephone line becomes a problem. .

On the other hand, the Captain System has two modes: (1) a mode in which the above image data is sent and received over a telephone line, and (2) a mode in which individual mosaic patterns are encoded and sent and received in the same way as characters. By generating image data by combining a plurality of mosaic patterns and transmitting/receiving the 2-1 conversion information of the mosaic pattern in the mode (2) above, it is possible to improve the usage efficiency of the telephone line. .

Conventionally, when converting image data into a mosaic coat, for example, a plurality of mosaic patterns stored in a mosaic pattern dictionary are read out and displayed on a display device, and a mosaic pattern that best suits the image data is selected in a so-called conversational format. Since the method of selecting patterns was used, there was a problem that it took a lot of time, and an effective method for mosaic coding of image data was awaited.

[Conventional technology]

FIG. 6 is a block diagram showing a conventional mosaic encoding device, in which 1 is a processor (hereinafter referred to as CPU), and 2 is a display device (hereinafter referred to as CRT).

3 is a keyboard, 4 is a random access memory (hereinafter referred to as RIM), and 5 is a file memory (hereinafter referred to as FD).
).

FIG. 7 shows an example of the display screen of the CRT 21 in the operation saic coding device when image data is mosaic coded using the mosaic coding device described above. The mosaic selection pattern ■ indicates the cursor.

Hereinafter, a conventional method of creating an image using a mosaic pattern will be explained with reference to FIG. 6, with reference to the flowchart of FIG. 7 and the display screen of FIG. 8.

First, by turning on the power to this mosaic encoding device, the operating system (O3) is transferred from the FD 5 to the RAM 4. The image creation program 41 etc. are initial program loaded (IPL), and when the initial program loading (IPL) is completed, the CPU
1, the image creation program 41 is executed, and a screen for creating an image by combining mosaic patterns is displayed on the CRT 20.

When the operator presses the image creation key on the keyboard 3, CI? A cursor ■ is displayed at a specific position (for example, on the left) on the screen of T2-1-.

Next, the operator moves the cursor (2) to the position (Kl, 2) where he wants to create an image. (Figure 7, step 50
．． (See Figure 8, Kl, 2) Next, when you press the mosaic display key on the keyboard 3', multiple mosaic patterns with numbers added to each mosaic pattern will be displayed as shown in Figure 8 (■). The selected pattern will be displayed. (See Step 51 in Figure 7) At this stage, the operator selects a mosaic pattern that matches the image ■ that he is about to draw, and enters the corresponding number.
For example, input using the numeric keypad on keyboard 3. (7th
(See steps 52 and 54 in the figure) If a matching pattern cannot be found among the displayed mosaic selection patterns ■, press a specific key on the keyboard 3,
Instructs to change the mosaic selection pattern display. (Figure 7,
(See step 53) In this way, when the mosaic pattern number suitable for image creation is input, the cursor position (Kl, 2) is input.
The selected mosaic pattern is displayed in ().

(See Figure 7, Step 55. Figure 8, Kl, 2) If you have confirmed that the mosaic pattern is most appropriate for the screen you want to create, move the cursor ■ to the next position (K3). Then, the process returns to step 50 to repeat the same operation, but if the selected mosaic pattern is inappropriate, the process returns to step 51 and the mosaic selection key is pressed. (See step 56 in FIG. 7) [Problem to be solved by the invention] As described above, in the conventional method, mosaic patterns are combined in a so-called conversational manner while checking the compatibility with the image to be drawn. Since images were generated, there was a problem in that it took a long time (for example, several hours) to create one image.

In view of the above-mentioned drawbacks of the conventional art, the present invention automatically selects a mosaic pattern suitable for the image to be drawn by computer processing without human intervention, and creates a mosaic near image data.
There is a method C whose purpose is to provide a method for unification.

[Means for solving problems]

The eleventh point is to quantize the image data from the image data input device and store it in the image memory, and then set a frame for image compression of n x m l' soto (n, m are positive integers).
a first means for reading image data within the frame;
A mosaic pattern read from a mosaic pattern dictionary that stores a plurality of mosaic patterns each consisting of m dots is compared with the image data read out by means of storage 1, and a Mori 2 pattern with the minimum degree of mismatch is found. 1
One or more mosaic patterns are extracted, and one of the extracted mosaic patterns is i! The present invention is achieved by the mosaic coating apparatus of the present invention, which further comprises a second means for generating a code, and converts the image data described in - into a mosaic code using the mosaic pattern obtained by the first means and the second means. Ru.

[Effect]

That is, according to the present invention, the original image is divided into, for example, one line 204.
Scanning is performed M times with an 8-dot (bit) image scanner, and the analog information for each pixel of the obtained 2048 x M dots is quantized by a binarization circuit and stored in an image memory. ×m dot (n, m
is a positive integer), cut out the frame information for image compression (encoding) that has the same shape as the mosaic pattern, and use the same nxm in the mosaic pattern dictionary stored in RAMζ.
The n x m l' sort stored in the image memory is used to select the mosaic pattern closest to the cut out image data by sequentially comparing and comparing all of the individual mosaic patterns of dots. By performing this on all frame data, the image data is converted into a mosaic code, so the effect is that the mosaic code on the original image data can be performed at high speed without human intervention. be.

〔Example〕

FIG. 1 is a block diagram showing an embodiment of the present invention, where the same symbols as in FIG. 6 indicate the same objects, and 7.8 are functional blocks necessary to implement the present invention. 7 is an image scanner, and 8 is an image memory (hereinafter referred to as GM). In order to carry out the present invention, in addition to this, an image data cutting program 42. It is also necessary to prepare a verification program 43.

FIG. 2 is a diagram explaining the present invention in detail, and FIG. 3 is a diagram explaining the present invention in detail.
Xm) This is a diagram explaining the concept of comparing image data of Soto with a mosaic pattern, FIG. 4 is a diagram showing the structure of a mosaic pattern dictionary, and FIG. It is a diagram schematically shown.

First, the present invention will be explained in detail with reference to FIG. 2 while referring to FIG.

When the CPU 1 executes the image data cutting program 42, the following processing is performed.

That is, an image drawn on a medium is scanned M times by an image scanner 7 in which the number of pixels per line is, for example, 2048 dots, and a binarization circuit (not shown) converts the image into O/1 binarized data. Quantize it and store it in GM8. (See “a”) Next, the above 22O48 on the GMB
An image compression (encoding) frame 81 consisting of n×m dots (8×12 dots in this example) is set on the image data of , stored in specific area 1 of RAM 41.

“b” indicates this state. In this figure,
The shaded area (black) indicates logic °l゛, and the white area indicates logic °0
° is shown.

Subsequently, by executing the verification program 43, the following processing is performed. That is, take out the mosaic pattern dictionary “C” from FD5 and
After storing it in the specific area J of AM 4, it sequentially reads out the mosaic pattern of n x m dots (8 x 12 dots in this example) at the start address, and reads the image data “b” stored at address 1 above. '' in pixel (dot) units, selects the mosaic pattern with the smallest degree of mismatch [indicated by "d"], and sets its mosaic pattern number as the mosaic coating number of the image data.

In the example in this figure, the mosaic pattern dictionary "
The mosaic pattern "e" numbered 24 from "C" is translated into iI, and its output code is "24"(").

By performing such processing on all the image compression frames 81 of nXm tosoto set in the above-mentioned image data 1- stored in GM8, mosaic coding of the image data is performed. become.

An example of the matching method shown in "d" above will be explained using FIGS. 3 to 5.

This method is generally known as the pattern matching method, and is used for character recognition in optical character recognition (OCR). For example, each dot in a mosaic pattern is assigned a weight (importance) of 0 to 3. In this method, the degree of mismatch is calculated by multiplying the dots by -1- weights in correspondence with the image data to be compared.

For the above weights, respectively 0: You can ignore it.

1: Weight is low.

2; Weight is normal.

3: High weight.

For example, for dots near the boundaries between black and white in each mosaic pattern, the weight is lowered there, and the weight is set higher at the periphery. By weighting in this way, highly accurate matching can be performed.

In Figure 3, A is image data and °O゛ is white.

1 indicates black. 82 shows an example of weighting of the mosaic pattern dictionary, where the - sign indicates white and the 10 sign indicates black. In this example, for convenience of explanation, a mosaic pattern of 4×6 dots is shown.

Now, when image data 8 is compared with mosaic pattern 81, among the dots of mosaic pattern 81,
Since the dots marked with * do not match, it can be seen from the respective weights that the degree of mismatch is I+1+3.5.

Next, when comparing image data A and mosaic pattern B2, in the same way, the degree of mismatch is -3+3+
It can be seen that it is 1+2+1+3+1.14.

Therefore, in this case, the old mosaic pattern with the smaller degree of mismatch is selected and its code number is used as the coat of the image data A.

FIG. 4 shows an example of a mosaic pattern dictionary "C", where C1 is a dictionary pattern, C2 and C3 are mask patterns, and each dot corresponding to the one stored in the two mask patterns is shown in FIG. 2'(n
・0. It is configured to indicate the weight indicated by l).

Therefore, by setting the mask pattern C3 to 2° (n=0) and setting the mask pattern C2 to 2° (n・1),
The weights of 0 to 3 can be known from the 0/1 information stored corresponding to each dot of the mask patterns C3 and C2.

A specific method for calculating the mismatch degree explained in FIG. 3 from the O/1 information corresponding to dots stored in the dictionary pattern C1 and mask patterns C2 and C3 will be explained with reference to FIG. First, from the image data stored in fllGMB, the data within the 8x12 dot frame is stored in the specific area I of RAM 4, and the data of that 1 line (1 byte - 8 dots) is stored in RAM 4. Store at address G.

RAM mosaic pattern dictionary C1-C3 from FD 5
For the dictionary pattern C1 with the mosaic pattern number "O", the first line data (1 byte = 8 dots) is stored in the D address of the RAM 4. Data of the same first line of mask patterns C2 and C3 corresponding to dictionary pattern CI are stored at addresses M1 and M2, respectively.

(See Figure 5, old P) (21CPU 1 reads the line data (G) and (B) set in the above RAM 4, and
(b) Perform the calculation =6G. (See Figure 5, PM2) (3) Next, CI"Ll l is calculated by performing the operation (G)・(Ml)-5 old" and counting the number of "1" for (old). It is stored at the old °° address, and the (old °') is shifted to the left by 1 bit, that is, the operation (x2) is performed. (Figure 5, P
(See M3) (4) Next, CPII l performs the operation (G)·(M2)→M2', counts the number of °l' for (M2'), and stores it at address M2'. (See Figure 5, PM4)
(5) Finally, the CPU 1 can calculate the degree of mismatch for the one line data by performing the following calculation: (old')+(M2') ->PM. (Refer to Figure 5, PM5) Therefore, the calculations in (1) to (5) above are performed as follows: -1 2 x 8 x]
By repeating this process for each line of 2-dot image data (12 lines in total) and summing the degree of inconsistency for each line, it is possible to obtain the degree of inconsistency of the image data with respect to mosaic pattern number 00.

The mismatch calculation operation is performed for all mosaic patterns, and the number of the mosaic pattern with the smallest mismatch is selected, and that number is used as the mosaic code for the image data of the 8X]2F throat.

By performing the above pattern matching operation on each piece of data within the 8×12 dot frame in the image data stored in the GM 8, the mosaic encoding of the image data is completed.

In the process of calculating the degree of mismatch described above, if multiple mosaic patterns with the smallest degree of mismatch are obtained for one image data of 8 × 12 dots, for example, the mosaic pattern with the smallest number is selected. , may be used as a mosaic code for the image data, but in general,
It is clear that there is no need to be particularly limited since the procedure boils down to selecting one from a plurality of the same items.

The pattern matching method described in detail above, in which a mosaic pattern closest to the image data of 8×12 dots is selected from the mosaic pattern dictionary, is one example, and is not based on the gist of the present invention. It goes without saying that this method is not limited to this method,
One of the main points of the present invention is to set an image compression frame of, for example, 8 x 12 dots, which is equal to the size of the mosaic pattern, to the image data (for example, 2048 x M dots) stored in the GM8. , and the other main focus of the present invention is to extract the data within the frame from the mosaic pattern dictionary for the frame data for image compression extracted from the image data. A close mosaic pattern is selected and the mosaic pattern number is used as a code for the image data within the frame.

〔Effect of the invention〕

As described in detail above, the mosaic coding device of the present invention converts an original image into, for example, 2048 dots per line (
Scanned M times with an image scanner (Focus), the obtained 2
Analog information about each pixel of the 2O48X dot is quantized by a binarization circuit and stored in an image memory. Then, information within the frame of the nxm dot in the same shape as the mosaic pattern is cut out from the image memory and stored in the RAM. The image memory is programmed to select the mosaic pattern closest to the cut out image data by sequentially comparing and comparing all the individual mosaic patterns of the same nXm dots in the mosaic pattern dictionary. Stored n
By doing this for all frame data of xm dot,
Since the image data is converted into a mosaic code, the original image data can be converted into a mosaic code at high speed without manual intervention.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing one embodiment of the present invention. FIG. 2 is a diagram illustrating the present invention in detail. FIG. 3 is a diagram explaining the concept of comparing nXm dot image data and a mosaic pattern. FIG. 4 is a diagram showing the structure of a mosaic pattern dictionary. FIG. 5 is a diagram schematically showing the matching operation in FIG. 3. FIG. 6 is a block diagram showing a conventional mosaic coding device. FIG. 7 is a flowchart showing the operation when performing mosaic coating of an image take using a conventional method. FIG. 8 is a diagram showing an example of a display screen on a display device in a conventional mosaic coating apparatus. In the drawings, 1 is a processor (CPU), 2 is a display device (CRT), and 2 is a display device (CRT).
). 3 is the keyboard. 4 is random access memory (RAM). 40 is an operating system (O3). 41 is an image creation progera 14. 42 is an image data cutting program. 43 is a verification program. 1. J is the memory area. G, D, old, old °, old", M2. M2" is the memory address. 5 is a file memory (FD). 7 is an image scanner, and 8 is an image memory (GM). Items are mosaic selection patterns. CI is a dictionary pattern, C2 and C3 are mask patterns. are shown respectively.

Claims (1)

[Claims] After quantizing the image data from the image data input device and storing it in the image memory, a frame for image compression of n×m dots (n and m are positive integers) is set, and the image data within the frame is read out. The first means compares the mosaic pattern read from a mosaic pattern dictionary storing a plurality of mosaic patterns consisting of the same n×m dots with the image data read by the first means, and there is no discrepancy. a second means for extracting one or more mosaic patterns with a minimum degree and selecting one from the extracted mosaic patterns; 1. A mosaic coding device for converting the image data into a mosaic code using a generated mosaic pattern.