JPH03127554A - Picture coder - Google Patents

Abstract

PURPOSE:To attain high speed processing by applying 2 kinds of signals at change points from white to black level and from black to white level as to a reference line and a signal representing a change point as to a coding line to plural stages of F/F and comparing outputs of the FFs so as to decide the coding mode of an MR code. CONSTITUTION:An AND gate 103 forms a change point signal caused with logical 1 of a Q output of a 1-bit D flip-flop (D-F/F) 101 and an inverse of an input picture element signal R, that is, taken as logical 1 for only a change picture element from black to white picture element on a reference line and an AND gate 104 forms a change point signal caused with logical 0 of an inverse of Q output of the 1-bit D flip-flop (D-F/F) 101 and the input picture element signal R, that is, taken as logical 1 for only a change picture element from white to blank picture element. On the other hand, an exclusive OR gate 105 exclusively ORs a Q output of a (D-F/F) 102 and an input picture element signal C on a coding line to form a change point signal in which a change pictured element from white to black and from black to white level is taken as logical 1. Then the three change point signals as above are compared to determine the MR coding mode. Thus, the high speed processing is attained.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an image encoding device for image signals used in facsimile machines, image file devices, and the like.

[Conventional technology]

Conventionally, in facsimile machines etc., MR (MR) is applied to image signals.
Modified READ) When performing encoding processing, the starting pixel of the encoding line is a. Binding, a.

The first point of change to the opposite color thereafter is al, and similarly on the reference line, the starting pixel a. The pixel directly above the

Binding, first change point b1 to the opposite color thereafter, a after bI. Let b2 be the point of change to the same color as . And vertically,
In order to determine the three types of encoding modes (horizontal and bus), distances aOaI+bob, , bob2 were counted, and then these counting results were compared and arithmetic operations were performed.

``Problem to be Solved by the Invention 3 As described above, in the conventional example, in order to determine the encoding mode of the MR code, a plurality of counters, a comparator for calculating the difference between the count results, and the following In preparation for encoding, these count results are processed according to the code mode,
Since multiple steps of memorization are required, these processes were performed under program control using a microcomputer.

Furthermore, it takes at least several clocks to determine the encoding mode for l change points, and as the frequency of occurrence of change points increases, the processing speed also slows down accordingly.

[Means to solve the problem]

The present invention has been made in view of the following two points, and aims to encode an image at high speed without performing complicated arithmetic operations. a first signal representing a change point from black to white and a second signal representing a change point from black to white; 2
a transmission means, a detection means for detecting whether a change point on a reference line has passed through a pixel position to be encoded, a counting means for counting the number of consecutive pixels of the same color on an encoding line, and a base point for encoding. The first . selection means for selecting one of the second signals, an output of the detection means, an output of the selection means and the second signal;
The present invention provides an image encoding device having a third signal transmitted by the transmitting means and a determining means for determining the encoding mode based on the third signal transmitted by the transmitting means.

〔Example〕

The present invention will be explained below using preferred embodiments.

First, before discussing the specific configuration of the embodiment, a brief explanation will be given of the modified READ (MR) code used in facsimile machines and the like.

In CCITT's Recommendation T4 regarding MR codes, changing points during two-dimensional encoding are defined as follows.

a(, - coding starting pixel a1 of the coding line - a above.

The first change point a on the right side of 2--The first change point on the right side of a1 above.-a above. The first change point on the reference line to the right and having the opposite color to aQ, b2 - the first change point on the reference line to the right of above, a thus defined. , al, a2 are on the encoding line, and the same <b, and b2 are on the reference line. And these a. + a l r a According to the relative position (distance) of the group 2 and the group b2,
It is specified that the encoding mode is uniquely selected from the following three modes and encoding is performed.

(+) Bus mode (P mode) When b2 is to the left of al (there is only one generation code) (2) Vertical mode (V mode) When albl l≦3 (a total of 7 generation codes differ depending on the distance) Seed occurrence code) (3) Horizontal mode (H mode) Above (1) (2) When outside B (according to the run length code table) Format: H + M (a Oa 1) +M (a + a
2) Here, H is a code indicating H mode, M(a.

a+) is a run-length code of 1 aoa, 1 for white or black, and M(ala2) is a run-length code of 1 a 1 a 2 for black or white.

However, if two or more of the above modes (1), (2), and (3) are satisfied at the same time, (1) P mode > (2) V mode > (3) H
Priority is given in order of mode.

FIG. 4 is a diagram explaining the basic operation of this embodiment.

FIG. 4(a) shows the starting pixel a. This shows the actual positional relationship of pixel data, taking as an example the encoding when the pixel is white.

C and R are changed pixel signals, C is coding line data in which changed pixels in the coding line are represented by 1 and other pixels are represented by O, and C1 represents the pixel to be encoded,
When co is 1, it indicates that the pixel to be encoded is a change point. In addition, R is reference line data in which the pixel at the point of change from white to black on the reference line is represented by 1, and other pixels are represented by 0, and R, represents the reference pixel immediately above C1, and Rn-3, Rn+3 indicates that they are 3 pixels to the left and 3 pixels to the right of Rn, respectively.

Also, RBI is data that becomes 1 when R1 becomes 1, and this state is maintained, and Rtar is Rn-3.
This data indicates that 1 (that is, the point of change from white to black) exists on the left side.

On the other hand, contrary to R1-3 to Rn+3, RB2 is data that represents the point of change from black to white in the reference line with 1, and indicates a pixel immediately above Cn.

Now, if we shift the above-mentioned change pixel signals C and R of the reference line and the encoded line in the state shown in FIG. 4(b) from the left to the right of the scanning line, the horizontal, vertical,
Each mode of the bus is classified into FIGS. 4(b) to 4(h). At this time, Rtar is a change point, but with respect to Cn,
Indicates that the object was located more than 3 pixels away from the left, and
RBI indicates that the change point b1 has already passed directly above Cn.

FIG. 4(b) shows the case of VR(3) in vertical mode.

Since Rfar = O, Rn-3 = 1, R
It is determined that n-3 is the change point S, and at this time C,=1
, that is, al has been detected. Note that the * mark in the diagram indicates that its content is 0. 1 Indicates that either one is acceptable.

Similarly, FIG. 4 <c> shows VR(2), and (d) shows V
L (2), (e) hail V t, (3).

FIG. 4(f) shows the first case where the horizontal mode occurs, and before the change point S occurs at R1-3 to Rn+3,
,=1, that is, a is detected. Also, (g)
is the second case, which is a horizontal mote, and Rfar ""
1.

This is the case where bl is located on the left side of Cn21, that is, more than three pixels from al.

Fig. 4(h) shows the case where it becomes a pass mode, and RIn=
l, Cr+=0. RB2:I, that is, it indicates that before detecting al, b2 was detected.

FIG. 1, FIG. 2, and FIG. 3 are examples of circuits embodying the basic operation described above. Note that all flip-flops and counters in these diagrams are reset at the start of each line, and all operate synchronously using a common clock signal.
Not shown.

In FIG. 1, 101. 102 is a 1-bit D flip-flop (hereinafter abbreviated as F/F); 109. 1.10
111,. 112 is 4-bit D F/F, 115,1
16,117 are 2-bit D F/F, 118 to 122
is J-K F/F.

132, 133 are 2tol selectors (inverted output), 1
34 is a 2tol selector (non-inverted output).

In addition, the input signal R is a binary serial pixel signal of the reference line, the input signal C is a binary serial pixel signal of the encoding line,
The input signal A is a signal indicating the end position of one line, which becomes 1 from the start of the l line and becomes O from the end pixel onward.The input signal DET is a signal which becomes 0 when the code mode is determined.

The output signal C0LOR is the encoding starting pixel a. Output signal END is a signal indicating that encoding has been completed up to the end of one line. Output signal EN is a signal that allows the run length counter in Fig. 3 to count. Output signal C1
Kiyokuchi to Kiyoshita, drunken tax Bl, and B2 are signals corresponding to each pixel signal shown in FIG. 4.

Note that the binary pixel signal is 0 for a white pixel and 1 for a black pixel.

The operation will be explained below.

D-F/FIO1, 102 are input pixel signals R, respectively.
The state of one pixel before 1C is maintained.

When the Q output of D-F/F1ol is 1 and the inverted signal of the input pixel signal R is I, the AND gate 103 selects 1 only for pixels that change from black pixels to white pixels on the reference line.
And, AND gate 104 is q of DF/FIOI
When the output is O and the input pixel signal R is 1, that is, only the pixels that change from white pixels to black pixels on the reference line are set to 1.
Create a changing point signal.

On the other hand, for the pixels of the encoded line, exclusive OR
K-1-105i: Te D-F/F1.02 By taking the exclusive OR of the Q output and the input pixel signal C, a change point signal is obtained in which a pixel that changes from white to black or from black to white is set to 1. make.

Next, the OR gates 106, 107, and 108 are set to 1 from the start of one line and become O after the end pixel.The end position display signal A forces the change point signal after the end pixel to 1 regardless of the input pixel signal. This is for the purpose of This satisfies the MR encoding rule of "placing a virtual change point immediately after the end pixel."

The above three change point signals and the end position display signal A, which is 1 only in the effective pixel range of one line, are 4 pins) 11)
- Sequentially transmitted by F/Fs 109 to 112.

Now, if the pixel to be encoded is C1 which is the Q3 output of the D-F/F 112, the outputs Ql and Q2 of the F/F 109 are 3 on the main scanning line with respect to the above pixel to be encoded Cn on the reference line. This indicates the state of the pixel located to the right of the pixel. Similarly, the Ql and Q2 outputs of F/FII0, 111°11, and 2 indicate the states of the second pixel, the first pixel, and the pixel on the reference line immediately above the pixel to be encoded, respectively, on the right side of the pixel to be encoded. Further, the output Q4 of the F/F 112 becomes a count enable signal for the run length counter.

On the other hand, the 2-bit DF/Fs 115, 116, and 117 sequentially transmit the outputs Ql and Q2 of the 4-bit DF/F 112. The state of the second and third pixels is shown. These F/F115. 116. 117
AND gates 124 to 129 at the inputs of these F/Fs 115 when the encoding mode is determined.
.

This is a gate to clear all the change point information in 116 and 117.

J'-K F/F118, 119 are 4-bit D-F/
F1], the outputs Ql and Q2 of 1 are used as J inputs, and the change point b of the reference line passes through the pixel position to be encoded during the detection of 1 encoding mode (input to 4 bit D-F/F112) This is a gate that detects whether the

The J-KF/F 120 receives the output Q3 of the 4-bit DF/F 112 as an encoding starting pixel a. , and inverts white/black (Q output) every time a change point l appears in the signal Cn. Based on this color information, the change point information from white to black can be determined from the change point information of the above reference line.
Select whether to refer to the change point information from black to white.

J-, F/F], 21. , 122 is F
/F117 output Ql.

This is a gate that receives Q2 as an input and detects that there is a change point to the left of four or more pixels with respect to the change point signal Cn corresponding to the pixel of interest. Therefore, before Cn"1, that is, before detecting the change point a of the encoded line, the current F/F 121
When the output Q of becomes 0, vertical mode encoding is impossible.

132 to 134 are selector circuits that select one of the two types of reference line signals applied to the input and the B input, respectively, based on the color display signal COLOR output from the F/F 120.

AND gate 135 outputs Ql and Q of F/Fill.
This is an end detection circuit that outputs 1 when both 2 become 1. That is, normally, the change from white to black and the change from black to white do not occur at the same time, and both change point signals are set to 1 by the OR gates 106 and 107 only at the end of the l line. The output of this AND gate 135 becomes 1. Therefore, based on this, it is possible to detect the end of the l line. The J-KF/F 136 is for synchronizing with other circuits, and the Q output=l indicates that an EOL code should be created. Further, the inverted output Q is connected to the least significant bit load input of a run length counter shown in FIG. 3, which will be described later. As a result, the run length count value is the normal starting pixel a. However, as an exception, if the first run in the horizontal mode reaches the end of one line, the second run can be set to 0.

In the manner described above, signals corresponding to the signals shown in FIG. 4 are obtained by this circuit.

FIG. 2 shows change point information Cn, Rn-3 to Rn+3, Rfy, B obtained by the circuit shown in FIG.
This circuit determines which of the vertical VN1 and horizontal H1 bus P modes of the MR code should be encoded based on the MR code.

In the figure, 201 is a priority encoder, which outputs one of the seven vertical modes as a 3-bit code when encoding is to be performed in the vertical mode. J-KF/F202
is for prohibiting other encoding modes until the two run lengths encoded by this are completed when the horizontal mode is detected.

The condition for vertical mode is that the relative distance between the change points a, b1 is 3 pixels or less, so Rfar = L.

When Cn=1, this corresponds to one of Rn-3 to Rn+3 being O, and the output V2. A code corresponding to the mode is output to V, , Vo. 203 is r; and CI+
This is a NAND gate for detecting that both are 1.

There are two conditions for the horizontal mode: when bl is 4 or more pixels on the left side of the change point a1, and when bl is 4 pixels or more on the right side.The former is r; 20, and the latter is lτfar =
1, Rn-3 to Rn+3 = 1, and when the output of the AND gate 204 becomes 1, it is determined that the mode is the horizontal mode.

In the case of pass mode, the change point al is detected without detecting the change point al, b2 is detected, and in this circuit, C
n=0. B1=1. There are times when the condition B2=1 is satisfied, and setting the output of the AND gate 205 to 1 indicates the pass mode.

The output signal DET is a signal indicating that some mode has been determined, and since the changing point of the encoding line is always encoded in some mode, first, there is a case where C1=1, and a case where C1=1.
When the pass mode is established, the output of the OR gate 206 is set to 1, and DET=1. Signal DE
T is a signal for initializing a predetermined portion in the circuits of FIGS. 1 and 3 each time one mode is determined.

FIG. 3 is an example of a run length counter used for horizontal mode encoding.

301 to 304 are 4-hit binary synchronous counters, the signal 1NiT is the q output of the JK F/F 136 in FIG. 1, the signal DET is the output of the inverter 207 in FIG. 2, and the signal EN is the D- in FIG. This is the Q4 output of F/F112.

The operation will be explained below.

First, at the start of encoding one line, the counter 30] to 30
41 is reset together with all the F/Fs shown in FIGS. 1 and 2. After that, the clock CK corresponding to the pixel input of the encoding line is counted, and 1 is added for each clock, but the input signal EN causes the first pixel of one line to be
The counting operation is inhibited until it appears at the Q3 output of the D-F/F112.

Further, each time the encoding mode is determined by the circuit shown in FIG. 2, an initial value is loaded into this counter by the input signal DET. As already mentioned, the initial value to be loaded is 1 in the middle of the line due to the signal 1NiT, and 0 after the end of the 1 line.

The basic operation is as described above, and the AND gate 305
Therefore, the 16-bit counter of this embodiment is OA 3 F
The next number after H is counted as 1040H. Thereafter, when the lower 12 bits become A3FH, the lowest 4 bits are incremented by 1, and the lower 12 bits are initialized to 040H.

With this configuration, CC1TT Recommendation T4
The maximum value of the makeup run length is 2560 pixels, and to represent a run longer than this, 2560 pixels is required.
Following the extended makeup code, another makeup code is added. Therefore,
The lowest 6 bits of this counter (Q A of counter 301
−Q o , QA of counter 302;

QB) represents the terminated run length, the middle 6 bits (QC, QD of counter 302, Q^~QD of counter 303) represent the makeup run length from 64 to 2560, and the most significant 4 bits (counter 304
Q to QD) is the makeup ensemble length 2560
It represents the number of times.

Note that if you count with a simple binary counter instead of this configuration, you will have to divide the count value by AOOH.
By setting A3FI (not 3FH), when extended makeup RL is other than 0, makeup RL part is 0.
This can be prevented.

FIG. 5 is a timing chart of main signals in the embodiments shown in FIGS. 1, 2, and 3.

In Figure 5, CI (is a clock that drives all F/Fs and counters in the circuit, and 1 is a signal that goes low with one pulse at each start of one line, which causes all F/Fs and counters to drive.
/Ii counter is reset.

R and C are serial reference line signals and encoded line signals synchronized with the clock CK with black pixel = 1 and white pixel = 0, respectively, A is set to 1 by the above CLR, and after a predetermined length of one line. 0 is the termination instruction signal, Cn is the change point signal of the encoding line to be encoded, DET is the output signal indicating that the encoding mode has been determined, ■2 to ■o are the vertical mode, from l to The output signal indicates this with a code of 7, H is an output signal that becomes 1 when it is a horizontal mode, and when H = 1, the value of the counter output RL and COL
The output of the OR signal indicates the first run length in horizontal mode, and the second run length is DET=1. , V2
~Vo, H, is indicated by the value of RT- when P=O.

P is an output signal that becomes l during bus mode, RL is the count value of the counter shown in FIG. 3, and C0LOR is the starting pixel a.

E N D is an output signal indicating that an EOL code is required by setting it to 1 at the end of encoding one line.

The following is a step-by-step explanation.

First, the counter RL of the run length counter shown in the third diagram is initialized to O by the CLR signal. The second pixel, which is the first change point of the encoded line, is three pixels to the left of the change point, so at CK6, DET=1. V2～V
By setting o=1, V L (3) is indicated. After this,
The counter RL is initialized to 1, and C0LOR is inverted to 1.

Next, at CK 8, DET=1. H=1. RL22
COLOR=1, indicating that the mode is horizontal and that the first run I/··ngs is black 2. The second following this
The run length of is D E T =
1. .

RL=3. Since C0LOR=O, it becomes white pixel 3.

At CK 14, DET=1. P=1 indicates pass mode.

At the end of the l line, the end signal A becomes O, so
OR gate 1.07 which is a change point signal. 1.06゜1
All outputs of 08 become 1. This is equivalent to placing a virtual change point next to the end pixel, and at CK=t,
DET=1. ■Vertical mode V by setting 2 to Vo=4. It is shown that Also, at the same time, END
The signal also becomes 1, indicating that one line ends with this code.

The above embodiment has been described for two-dimensional encoding of MR code, but it can also be applied to ~i MR encoding, and if the image data input as a reference line is fixed to 0,
By inhibiting the output of each of the , H, and P signals, this method can also be applied to one-dimensional encoding.

Further, as with the reference line, two types of signals indicating the change points from white to black and from black to white are used as the signals of the encoded line, and the starting pixel a is determined. A similar effect can be obtained by selecting the color according to the color. Further, at this time, as a condition for detecting the end of one line, it is also possible to use a condition in which both of the two types of signals on the encoded line side are l.

Another method for detecting the end of one line is the F/
It can also be set as the time when the output Q-1 of F112 changes from 1 to 0.

As explained above, the reference line has two types of signals: the point of change from white to black and the point of change from black to white, and the sign (
For the hill line, the signal representing the change point is transmitted through multiple stages of F/
The encoding mode of the MR code can be determined simply by transmitting F and comparing their outputs without performing any arithmetic operations, making it possible to perform extremely high-speed processing.

Furthermore, by setting all of the three types of change point signals to 1 after the last pixel of the l line, a virtual change point necessary for encoding can be set, and the two types of change point signals of the reference line are set to 1 at the same time. This means that the end of the l line is reached when the line is reached, and the end condition for one line can be set using a chain wheel method.

Furthermore, by configuring the run length counter as described above, when the run length exceeds 2560, there is no need to perform an arithmetic operation on how many make-up run lengths should be used to combine the count results of the run lengths.

〔Effect of the invention〕

As explained above, according to the present invention, the first signal representing the change point from white to black of the pixels of the reference line and the second signal representing the change point from black to white are transmitted. means, second transmission means for transmitting a third signal representing a point of change in color of a pixel on the encoding line, and detection means for detecting whether or not the point of change on the reference line has passed through a pixel position to be encoded. , a counting means for counting the consecutive number of pixels of the same color in the encoding line, and selecting one of the first and second signals transmitted by the first transmitting means according to the color of the base pixel of encoding. and determining means for determining the encoding mode based on the output of the detection means, the output of the selection means, and the third signal transmitted by the second transmission means, so that complex arithmetic operations can be performed. This makes it possible to perform high-speed encoding without any problems.

[Brief explanation of the drawing]

Fig. 1 is a circuit diagram for transmitting the changing point signal of the embodiment, Fig. 2 is a circuit diagram for determining the code mode from the changing point information shown in Fig. 1, and Fig. 3 is a diagram showing a counter for counting the run length. , FIG. 4 is a schematic diagram showing the concept of the present invention, and FIG. 5 is a timing chart of signals obtained by the circuits of FIGS. 1, 2, and 3 of the above embodiment. 103 104 05 106, 107. 1.09～112 132～134 135 ・・・・・・・・・・・・・・・・・・・・・
・・AND gate exclusive OR gate 108 ・・・・・・・・・・・・・・・・・・・・
・・・・・・OR Kate 115-122 ・・・・・・・・・
Flippuff mouth selector~no

Claims (1)

[Scope of Claims] (1) A first transmission means for transmitting a first signal representing a point of change from white to black of a pixel on a reference line and a second signal representing a point of change from black to white; and encoding. a second transmission means for transmitting a third signal representing a color change point of a pixel on the line; a detection means for detecting whether a change point on the reference line passes through a pixel position to be encoded; a counting means for counting the number of consecutive pixels of the same color; a selection means for selecting one of the first and second signals transmitted by the first transmission means according to the color of the base pixel for encoding; An image encoding device characterized by comprising: determining means for determining an encoding mode based on the output of the detecting means, the output of the selecting means, and a third signal transmitted by the second transmitting means. (2) In claim 1, after the end of one line, the first and second transmitting means indicate a change point.
, a second signal, and a third signal. (3) The image encoding apparatus according to claim 2, wherein the end of one line is determined when both the first and second signals transmitted by the first transmitting means represent a changing point.