JP3147246B2 - Image processing apparatus and method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus and method,
More specifically, the present invention relates to an image processing apparatus and method for inputting image data and generating an output image.

[0002]

2. Description of the Related Art Conventionally, in an image processing apparatus such as a facsimile apparatus, an image is read by an image scanner, and each pixel of the raster-scanned image is represented by binary data of white or black. When communicating the binary raster image data, the binary raster image is encoded and compressed by a method such as MH, MR, MMR or the like, and is transmitted and received. Therefore, in a conventional facsimile apparatus, as an internal representation of an image, image data is treated as a set of binary raster pixel data, and an image read in G3 fine (main scanning 8 pel / mm, sub-scanning 7.7 line / mm) is used. , Standard (main scanning 8 pel / mm, sub scanning 3.8
When the resolution of the image read on the transmission side is different from the resolution of the reception side, such as when transmitting to a receiver having only a recording resolution of 5 lines / mm), or a B4 size received image is output on A4 recording paper. If the received image and the output image have different paper sizes, the MH, MR, and MMR encoded image data is temporarily converted to binary raster image data, and is simply written twice for each pixel (the same pixel is written twice or more). (Continue) and scaling processing such as enlargement and reduction by thinning processing.

[0003]

However, in the above-described conventional example, when low-resolution image data is recorded on a high-resolution one, such as when a 200 dpi image is recorded on a 400 dpi recording paper, a simple pixel-by-pixel method is used. In order to perform the overlap processing, the jagged edges of the oblique lines are noticeable.
In the resolution conversion near 200% for converting to an image with a resolution of 00 dpi, there is a problem that the image is distorted because the resolution conversion is performed by the overlap thinning process for each pixel.

[0004]

SUMMARY OF THE INVENTION The present applicant proposes to solve this problem by reading an edge vector of a character and line drawing image and processing the image based on the extracted edge data. You are. It has been found that a certain degree of discrimination is possible without specially preparing a dedicated circuit for discriminating between a character line drawing area (binary image area) and a halftone image area in a document image. The present invention has been made in view of such a point, and when performing a series of processes by extracting an edge vector of an image, it distinguishes between a halftone image and a binary image based on a vector obtained in the process. It is an object of the present invention to provide an image processing apparatus and an image processing method which make it possible. In order to solve this problem, the image processing apparatus of the present invention has the following configuration. That is, contour extraction means for extracting contour vector data along the edge of the input image, counting means for counting points representing vectors for each contour closed loop extracted by the contour extraction means, and an extracted contour vector. Detecting means for detecting the size of each closed loop based on the data, and an image area represented by each closed loop based on the size of each closed loop detected by the detecting means and the number of vectors corresponding to each closed loop. Determining means for determining whether the image is a halftone image or a character / line drawing image.

[0005] The image processing method of the present invention includes the following steps.

A contour extraction step for extracting contour vector data along an edge of an input image, a counting step for counting points representing a vector for each extracted contour closed loop,
A detection process for detecting the size of each closed loop based on the extracted contour vector data, and a closed loop based on the size of each closed loop detected in the detection process and the number of vector points corresponding to each of the closed loops, is represented by each closed loop. And a discrimination step for discriminating whether the image area to be processed is a halftone image or a character / line image.

A further object of the present invention is to provide an image processing apparatus capable of distinguishing between a binary image and a halftone image with a simple configuration and reproducing a good image. It is. In order to solve this problem, the image processing apparatus of the present invention has the following configuration. That is, contour extraction means for extracting contour vector data along the edge of the input image, smoothing means for smoothing the extracted contour vector data, and output image according to the contour vector data smoothed by the smoothing means. Generating means for generating a contour vector, and further comprising: counting means for counting points representing a vector for each contour closed loop extracted by the contour extracting means; and Detecting means for detecting the size of each closed loop, and energizing the smoothing processing means for each closed loop based on the size of each closed loop detected by the detecting means and the number of vectors corresponding to each closed loop. Control means for controlling whether or not to do so.

Further, the image processing method of the present invention has the following steps. That is, a contour extraction step for extracting contour vector data along the edge of the input image, a smoothing step for smoothing the extracted contour vector data, and an output image according to the contour vector data smoothed by the smoothing step. And a generation step of generating points for each of the contour closed loops extracted in the contour extraction step. Based on the size of each of the closed loops detected in the detection process and the number of vectors corresponding to each of the closed loops, and energizing the smoothing process for each of the closed loops. And a control step for controlling whether or not to perform.

[0009]

In the configuration or process of the present invention, for example, contour vector data along the edge of the input image is extracted, and the total score representing the extracted extracted vector in the closed loop is obtained. Then, based on the relationship with the size of the corresponding closed loop, it is determined whether the closed loop image of interest is a binary image or a halftone image.

[0010]

Embodiments of the present invention will be described below in detail with reference to the accompanying drawings.

<Description of Apparatus Configuration> FIG. 1 is a block diagram of an image processing communication apparatus according to an embodiment. In the figure, reference numeral 100 denotes an image scanner, which performs A / D conversion on image data read in raster scanning order by a photoelectric conversion device such as a CCD, and further converts the A / D converted digital multi-valued data into a binarizing circuit. , And is output to the bus A1500 as binary raster image data. Bus A
Is output to the memory A once.
300.

Reference numeral 200 denotes an image area separating circuit, which is raster scanning multi-valued data input from the image scanner 100 (for example, in the case of 256 gradations per pixel, 8 bits / pixel).
1), the halftone area and the character / line drawing area are separated, and an image area separation table (to be described in detail later) for each image area is created in the memory A300. Note that which part of the read image is a binary image such as a character or a line drawing and which part is a halftone image can be determined by checking the value of the input multi-valued data. That is, in general, a binary image literally has only two states of white (lowest density) and black (highest density). Therefore, the density change in an area (edge) where these two states are adjacent to each other is sharp. . On the other hand, since the density of the halftone image changes gradually, the density change becomes gentle. Therefore, by examining this density change, it becomes clear whether the pixel position of interest is at the edge of the binary image or at the edge of the halftone image.
A technique for distinguishing such a binary image such as a character or a line drawing from a halftone image is already known, and may be realized by other processing. Further, the operator may instruct the type of the document to be read and the like.

Reference numeral 700 denotes a contour extraction circuit which reads the binary raster image data stored in the memory A 300 and an image area separation table for image area separation if necessary, and
At 00, a contour coordinate data sequence is created.

An encoding / decoding circuit 800 reads binary raster image data stored in the memory A 300 and performs encoding and compression of a binary image such as MH, MR, and MMR when performing encoding processing. The encoded image data is stored in B1100. When performing the decoding process, the memory B
The encoded image data stored in 1100 is input, and the decoded binary raster image data is output to the memory A300.

Reference numeral 900 denotes a smoothing / coordinate conversion circuit, which inputs a contour coordinate data string stored in the memory B 1100, performs various coordinate conversions (multiplication of magnification, x-coordinate y-coordinate exchange), and contours during enlargement / reduction. Is performed, and a converted contour coordinate data sequence is created in the memory B1100.

Reference numeral 1300 denotes a communication control circuit (CCU) which inputs image data (contour coordinate data string or image coded data and, if necessary, an area separation table) stored in the memory B 1100, and transmits the image data to a communication line. Is output. The image data received from the communication line is stored in the memory B.
1100.

Numeral 1000 denotes a binary image regenerating circuit which inputs a contour coordinate data string stored in a memory B 1100, draws a contour in a memory A 300 based on the contour coordinate data, and generates a binary raster image (bit map). Image).

Reference numeral 400 denotes an intermediate coating circuit, which inputs a binary raster image (bitmap image) in which a contour created in the memory A 300 is drawn, fills the inside of the contour closed area, and stores it in the FIFO memory 500. And so on.

The FIFO memory 500 resynchronizes the horizontal and vertical synchronization with the printer and outputs binary raster image data to the printer 600.

Reference numeral 600 denotes a printer such as an LBP, FIFO5
At 00, the binary raster image data resynchronized is input and the image data is recorded on a recording sheet.

<Explanation of Processing Contents> Next, the flow of image data of the image processing communication apparatus according to the present embodiment will be described with reference to the flow chart of FIG. 2 for transmission operation, FIG. 3 for reception operation, and FIG. Each will be described in detail.

<Explanation of Transmission Process> First, the procedure of the transmission operation will be described with reference to the flowchart of FIG.

In the image reading at step S1, the image is read by the image scanner 100 as binary raster image data, transferred to the memory A300 and stored.
At this time, when transmission other than immediate transmission, that is, when image data transmission is not performed immediately, the memory A300 is used as shown in steps S9, S10, and S11 for effective use of the memory.
Is encoded and compressed by an encoding / decoding circuit 800 and stored as an image file in a memory B1100, and is decoded at the time of transmission to be developed as a binary raster image in a memory A300. Further, the multi-value raster image data from the image scanner 100 is output to the image area separation circuit 200 simultaneously with the accumulation of the binary raster image. The image area separation circuit 200 determines whether the image is a character / line image area or a halftone area for each block (predetermined area), and creates an area determination table indicating the properties of the image area in the memory A300. Of course, the stored encoded image data file and the area determination table file have the same file attribute, and are extracted when necessary.

At this time, it is clear whether or not pseudo halftone is used by pressing a selection button (halftone button provided on an operation unit (not shown)) for performing pseudo halftone processing at the time of binarization. In this case, the image area separation table for each image area need not be created, and only the image attributes of one page such as pseudo halftone or character / line drawing need be recorded.

In step S3, an area determination table indicating whether the area is a character / line drawing area or a pseudo halftone area is stored in the memory A.
300, and if it is a character / line drawing area, step S
In step S4, the process branches to step S12.
If it is determined that the image is in the halftone area, and the process proceeds to step S12, and if it is determined that zooming is necessary, 2 is determined in the same manner as the conventional G3 and G4.
After the scaling process for each raster, which is the scaling process for the value raster image,
The encoding is performed according to the communication line, and the image data is transmitted.

When transmitting the image area determination table at this time, the area determination table in the memory A 300 is transferred to the memory B 1100 and transmitted.

In step S4, if the receiving side has a function of regenerating a binary raster image from the outline coordinate data sequence as image data, and determines that the outline coordinate data sequence can be transmitted, it is determined that the receiving side has the function. Is notified to the transmitting side, and the process proceeds to step S5. If the receiving side has only the conventional G3 or G4 function and cannot receive the contour coordinate data sequence, the process proceeds to step S16.
In step S16, it is determined whether or not scaling is required. If necessary, the process proceeds to step S17, where scaling processing is performed using the outline coordinate data sequence. If no scaling is required, the process directly proceeds to step S23, where the data is transmitted after the encoding process.

In step S17, the binary raster image data stored in the memory A300 is input to the contour extracting circuit 700, and the binary raster image data is converted into a contour coordinate data sequence. Accumulate columns. In step S18, it is determined whether or not enlargement processing is necessary for the outline coordinate data sequence stored in the memory B1100. If enlargement processing is necessary, the process proceeds to step S19.

In step S19, a contour coordinate data string is input from the memory B1100, and a contour smoothing process described later is performed. Then, the flow advances to step S20. In step S20, each coordinate value is multiplied by a scaling factor to convert the coordinate value into a coordinate value according to the transmission resolution. In step S21, the converted outline coordinate data sequence is transferred from the memory B1100 to the binary image regeneration circuit 1000, and the outline is drawn on the bitmap image in the memory A300. In step S22, the image bitmap on which the outline is drawn in the middle-painting circuit 400 is fetched for each raster, and is developed in the memory A300 as binary raster image data on which the middle-painting is performed.

In step S23, encoding in accordance with the communication mode is performed by the encoding / decoding circuit 800 and transferred to the memory B1100. The encoded image data transferred to the memory B 1100 is the CCU 1300 in the case of G4.
Sent by MODEM not shown for G3
Sent by

The processing after step S5 will be described. If the image area to be transmitted is a character / line drawing and the receiving side can receive the contour coordinate data, the process proceeds to step S5. Contour extraction circuit 70 as character / line drawing area
The image data transferred from the memory A300 to "0" is scanned for all images while updating a coarse contour coordinate data table described later, and finally converted to a contour coordinate data sequence and stored in the memory B1100. In step S6, it is determined whether or not the image data converted into the outline coordinate data sequence needs to be scaled. If necessary, the process proceeds to step S7 where the memory B
The image data stored in 500 is transferred to the smoothing / coordinate conversion circuit 800, and the coordinate value is multiplied by a scaling factor. At this time, if necessary, the coordinate sequence is tracked to leave feature points in the outline coordinate data sequence, and smoothing conversion is performed in the enlargement process. Further, if it is determined that the image can be restored by smoothing by tracking the image, the number of significant digits of the coordinate value of the coordinate data is reduced to compress the data amount. After the processing means, the outline coordinate data sequence subjected to coordinate conversion by the smoothing / coordinate conversion circuit 900 is returned to the memory B1100. In step S8,
Image data whose transmission resolution and reception resolution match are CCU
600. In the case of G3, it is transmitted by a MODEM (not shown).

By the way, in consideration of the communication error of the image data, if the communication device is a G4 CCU, if the image is to be communicated as the outline coordinate data sequence, if the protocol error correction and error image retransmission are performed, the outline coordinate data sequence Can transmission and reception be possible, or in the case of G3 using MODEM, E
Data transmission / reception using HDLC for CM is required. Therefore, when transmitting and receiving the outline coordinate data string through the MODEM by G3 and using the conventional telephone line, the forced E
When the CM mode is set and the ECM retransmission frequency is high or the ECM transmission / reception cannot be performed due to the line condition, the image data is switched to the conventional image data transmission using the MH and MR codes, and the image communication is performed in the conventional G3 mode. I do.

<Description of Reception Processing> Next, the reception processing of this apparatus will be described with reference to the flowchart of FIG.

In the reception in step S101, the CCU 13
00 or image data is received from a MODEM (not shown). The received image is conventional coded image data such as MH and MR based on G3 and MMR based on G4, or a contour coordinate data string transmitted by the image processing communication device described in the embodiment. The received image data is temporarily stored in the memory B1100.

In step S102, it is determined whether or not the received image data is a contour coordinate data string transmitted from the image processing communication device.

If it is determined that the image data is received from the apparatus described in the embodiment, the flow branches to step S114 to receive the MH, MR, and MH received from the normal facsimile apparatus.
If it is determined that the data is MMR code data, step S103
Proceed to.

In step S114, it is determined whether or not the received image data needs to be scaled. If it is determined that the scaling process is necessary, the process proceeds to step S107, and thereafter, the scaling process using the outline coordinate data sequence is performed. If the scaling process is not necessary, the process proceeds to step S109 for determining the coincidence in the scanning direction.
It is returned to the value raster image data and printed out.

When the received image data is MH, MR, M
When it is determined that the received image data is MR-encoded image data, and the process proceeds to step S103, the received data is sent to the encoding / decoding circuit 800, where the received data is decoded into binary raster image data, and is stored in the memory A300. Is accumulated in In the next step S104, it is determined whether or not scaling is necessary when the received image data is printed out. If it is determined that the magnification is necessary, the process proceeds to step S105. If not, the process proceeds to step S113. A300 to FIFO500
The binary raster image data is transferred to the printer 600, and the printer 600 is re-synchronized with the binary raster image data and printed out from the printer 600.

In step S105, it is determined whether the received image data is a pseudo halftone image or a character / line image, based on area determination (image attribute or area determination table). Then, the process proceeds to step S116 for the pseudo halftone image data block, and the raster image scaling circuit 160
0 and is scaled as a binary raster image. Also,
The image block determined to be a character / line drawing area is transferred to the contour extraction circuit 700, and subjected to a scaling process as a contour coordinate data sequence described below.

In step S106, the binary raster image data of the character / line image stored in the memory A300 is transferred to the contour extracting circuit 700, converted from the binary raster image data into a row of contour coordinate data, and stored in the memory B1100. You.

In step S107, it is determined whether any of the sub-scanning and the main scanning is an enlargement process. If it is an enlargement process, the process proceeds to step S115, and the memory B110
The contour coordinate data sequence stored in the sequence No. 0 is transferred to the smoothing / coordinate conversion circuit 900, and a first smoothing process and a second smoothing process by tracking pattern matching of the contour coordinate data sequence described later are performed, and the contours other than the corner points are processed. Performs coordinate data smoothing. With this processing, jagged oblique lines at the time of enlargement are suppressed.

In step S108, the contour coordinate data string or the contour coordinate data string whose contour has been smoothed is converted into coordinate data corresponding to the resolution of the print output by multiplying each coordinate string by a scaling factor.

In step S109, it is determined whether or not the scanning direction of the image data is different from the image scanning direction at the time of recording, for example, when the image read in B5 portrait is output in A4 landscape. In this case, the process proceeds to step S110, where the x coordinate and the y coordinate of the outline coordinate data sequence are exchanged, and
The data is stored in the memory B1100. If there is no need to exchange coordinates, the coordinates are directly stored in the memory B1100.

In step S111, the memory B1100
Is input to the binary image regenerating circuit, and a part of the memory A300 is used for storing a bitmap image of an output image.
Draw the outline of the bitmap image.

In step S112, the image data on which the contour line is drawn is taken out from the memory A300 for each raster, and the inside of the contour is painted by the filling circuit 400 to obtain the FI.
Transfer to FO500. At this point, the outline coordinate data is converted again into binary raster image data. At this time, the binary raster image stored in the memory A300 by branching in step S104 or step S105 is also taken out from the memory A300 at this time and is also read out by F
The data is transferred to the IFO 500.

In step S1013, the binary raster image data transferred to the FIFO 500
Synchronized with 0, the image is output from the printer 600 as a raster image.

<Description of Copy Processing> Next, the contents of processing at the time of copying will be described with reference to the flowchart of FIG.

In step S201, a document image is read. In this case, if there is no need to accumulate the files as in the case of the immediate copy, steps S202 to S213
The image data output from the image scanner 100 is transferred to the FIFO 500 via the memory A300 and printed out. Here, the data is temporarily stored in the memory A300 in order to absorb the difference in speed between the image scanner and the printer. If the printing speed can be changed like a thermal printer, the data is directly transferred to the FIFO 500 as described above. And output it. In the case of immediate copying, if the reading resolution of the image scanner 100 is the same as the output resolution of the printer, there is no need to perform any processing, and the transfer processing described above may be performed.

If it is instructed not to copy immediately but to copy the image data once after storing it as a file, the process proceeds to step S203, where the image file is encoded and stored in order to increase the memory efficiency of the image file. . Here, it is assumed that the binary raster image stored in the memory A300 is transferred to the encoding / decoding circuit 800, converted into an MH code or the like, and transferred and stored in the memory B1100.
For the encoding compression, various compression methods other than the MH can be used, and a compression method unique to the present apparatus may be employed. The image data once accumulated is read out again from the memory B1100 at the time of the print output operation as shown in step S204, and the encoding / decoding circuit 800
And the data is expanded as binary raster image data in the memory A300.

In step S206, it is determined whether or not the stored image file needs to be scaled. The basis of the determination here is a case where the document image size is different from the size of the apparatus, a case where an operator instructs, and the like. If no scaling is required, the process proceeds to step S213,
The image data is transferred to the FIFO 500 on a raster-by-raster basis, synchronized with the printer 600, and printed out. If the scaling process is to be performed, the process proceeds to step S207, where it is determined whether the stored image file is a pseudo-halftone image or a character / line drawing image based on the area determination table as shown in the transmission process. I do. In the case of a pseudo halftone image, the decoded image is transferred to the raster image scaling circuit 1600,
The value is scaled in the state of the value raster image and returned to the bitmap image memory of the memory A300. If it is determined that the block is a character / line drawing image, the process proceeds to step S208, and the block determined to be a character / line drawing area is extracted by the contour extraction circuit 70.
0, sequentially converting the binary raster image data into a contour coordinate data string, and storing the converted data as a memory B1100 contour coordinate data string. As described in the previous reception processing, the standard resolution of G3 (main scanning 8 pel / mm, sub-scanning 3.85 pe)
1 / mm), the enlargement processing for outputting the image data accumulated at 400 dpi to a printer of 400 dpi is performed.
In order to perform the smoothing process at 214, the smoothing / coordinate conversion circuit performs first smoothing and second smoothing processes as described later.

In step S210, each coordinate value of the contour coordinate data string or the smoothed contour coordinate data string is multiplied by a scaling factor to convert the coordinate value into a contour coordinate data string matching the resolution and size of the printer output image. . In step S211, the image data converted to the outline coordinate data sequence of the printer output image is transferred to the binary image regenerating circuit 1000, and the binary image regenerating circuit 1000 draws an outline in the bitmap image memory of the memory A300. Is done. In step S212, the memory A3
The binary raster image data on which the contour line formed at 00 is drawn is read for each raster, transferred to the FIFO 500, and printed out in synchronization with the printer as shown in step S213.

The above processing is described step by step in each stage. Of course, the decoding processing, contour extraction processing, smoothing processing, coordinate conversion processing,
The binary image regenerating process and the intermediate coating process can be performed in parallel, and the system throughput can be increased. Further, the processing having a large load in the present embodiment may be performed during standby instead of during transmission / reception.

<Explanation of Each Processing Circuit> Next, in order to explain in detail the resolution conversion by the outline coordinate sequence in each processing block, an outline extraction circuit, a smoothing / coordinate conversion circuit, a binary image regenerating image circuit , The intermediate coating circuit will be described in detail with reference to the drawings.

<Contour Extraction Circuit> The contour extraction circuit 700
As shown in FIG. 6, the pixel of interest 611 in the binary image
And eight pixels (A, B, C, D, 0, 1,
The process proceeds in accordance with the state of (2) and (3). The target pixel is raster-scanned, and the entire image is sequentially processed while being shifted for each pixel. Specifically, it is as follows.

The starting point of the rough contour vector extracted here
The end point is located at an intermediate position between pixels in both the main scanning direction and the sub-scanning direction. In both the main scanning direction and the sub-scanning direction, the center position of a pixel is indicated by a positive integer, and the pixel position is represented by two-dimensional coordinates. For example, in an image of 1728 pixels in the main scanning direction and 2287 pixels in the sub-scanning direction, when the position of the pixel of interest is [3, 7] (meaning the third pixel position of the seventh raster), the rough contour surrounding the pixel The vector coordinate sequence is clockwise (2.5, 6.5),
(3.5, 6.5), (3.5, 7.5), (2.5,
7.5). That is, these are the coordinates of the four corners when the target pixel is rectangular. The rough contour vector in this case is such that the above four points are a, b, c, d, and the vector is x →
In the y format, a → b, b → c, c → d, d → a.

FIG. 5 is a diagram showing an internal block of the contour extraction circuit 700. Reference numeral 510 denotes an input control circuit which takes in the binary raster image data stored in the memory A300, outputs it as serial data. This data is output as serial data in synchronization with a horizontal and vertical synchronization signal (not shown). Reference numerals 521 and 522 denote FIFO memories for extracting pixel data over three lines in the vertical direction, and delay line data of serially input pixel data line by line. Reference numeral 530 denotes a 3 × 3 shift register group which receives an output from the input control circuit 510, an output delayed by one line from the FIFO 521, and an output delayed by two lines from the FIFO 522, and outputs data of three lines. Enter sequentially. Then, the input three lines of binary image data are shifted one pixel at a time,
The data of × 3 pixels is output to the decoder 540. The decoder 540 outputs 4-bit numerical data indicating the state of the surrounding eight pixels excluding the target pixel (here, x in FIG. 6) and the value of the target pixel to the contour extraction calculation circuit 570. The state of the surrounding eight pixels here is information indicating the state where the pixel of interest is placed. For example, whether the pixel of interest is at the edge of the binary image, This indicates which direction on the left and right is white (binary data is “0”).

Reference numeral 550 denotes a main scanning counter which outputs the pixel position of the target pixel in the main scanning direction to the contour extraction operation circuit 570. A sub-scanning counter 560 outputs the pixel position of interest in the main scanning direction to the contour extraction calculation circuit 570.

The contour extraction operation circuit 570 is provided for the decoder 54.
A coarse contour vector is extracted based on the output value from 0, and the coordinates of the starting point of each coarse contour vector and another coarse contour vector flowing into the coarse contour vector (the starting point of this vector is the end point) A table having the item number information of another coarse contour vector flowing out of the coarse contour vector (the starting point is the coordinates of the end point of this vector) is created and updated.

Here, the above-described vector will be described as follows.

In the embodiment, the vector along the edge of the binary image is extracted. However, when the vector is extracted in consideration of the pixel level, the vector can be divided into a vertical vector and a horizontal vector. In other words, the edge of the binary image is a set of vectors in the vertical and horizontal directions. In addition, when paying attention to a certain vertical vector, a horizontal vector is always connected to the start point and the end point of the vector, that is, the vertical vector is sandwiched between the horizontal vectors. Conversely, a horizontal vector is sandwiched between vertical vectors. In the embodiment, when a certain vector is noticed, a vector connected to the start point of the vector, that is, a vector in which the start point position of the notice vector is the end point position is called an inflow vector, and a vector having the end point of the notice vector as the start point Are called outflow vectors. Then, by obtaining the connection relation of the individual vectors in this way, it is possible to specify the contour along the edge of the target binary image. Each vector is provided with a unique number for specifying the vector, and the number clarifies the connection relationship between the vectors.
The number specifying each vector is called an item number.

FIG. 7 is a diagram showing a horizontal coarse contour vector, and FIG. 8 is a diagram showing a vertical coarse contour vector.
These are stored in the memory B1 as a table as described above.
100 is stored.

In the figure, reference numeral 710 denotes a counter indicating the item number of the horizontal coarse contour vector, and is incremented by one each time the horizontal coarse contour vector is increased by one (newly found). 7
21 is an area used to record the x coordinate of the starting point of the horizontal coarse contour vector, and 722 is an area used to record the y coordinate of the starting point of the horizontal coarse contour vector. Reference numeral 723 denotes an area for storing an item number of a vertical vector (outflow vector) connected to the start point of the horizontal vector, and reference numeral 724 denotes an item number of a vertical vector (outflow vector) to which the end point of the horizontal vector is connected. It is. Again, since the two vectors connected to the horizontal vector are vertical vectors, these areas 723 and 724 store the item numbers of the vertical vectors connected to the horizontal vector of interest.

This concept is the same for the vertical vector table shown in FIG. That is, a counter 810 indicates the item number of the vertical coarse contour vector, and is incremented by one every time the vertical coarse contour vector is increased by one. An area 821 is used to record the x coordinate of the starting point of the vertical coarse contour vector, and an area 822 is used to record the y coordinate of the starting point of the vertical coarse contour vector. 823 is the item number of the horizontal vector (inflow vector) connected to the start point of this vertical vector, 824
Is an area used to record the item number of the horizontal vector (outflow vector) to which the end point of this vertical vector is connected.

In these horizontal and vertical coarse contour vector tables, a coarse contour vector is added and the contents of each table are updated as needed every time one pixel is processed. And a rough outline vector table in the vertical direction.

In this manner, by tracing the item numbers of the inflow and both output vectors of the horizontal and vertical coarse contour vector tables shown in FIGS. 7 and 8, the total number of contour lines in the image as shown in FIG. A table of a coarse contour vector coordinate sequence representing the total number of contour lines for each line (contour closed loop) and the x-coordinate and y-coordinate of each point in the contour is created, and a series of processing ends.

In order to improve the processing speed, if it is determined that the target pixel position is not at the edge of the image, the table is not substantially updated. -The update process may not be performed.

Here, a specific example of the process of creating / updating the above-described horizontal / vertical coarse contour vector table will be described. For the sake of simplicity, it is assumed that the read image has a size of 6 × 5 pixels, and a single pixel A and continuous pixels B and C exist as shown in FIG. Here, the coordinate positions (coarse contour vector coordinates) of the four corners of the pixel A are (2.5, 2..
5), (2.5, 3.5), (1.5, 3.5),
(1.5, 2.5). The pixels B and C need not be described.

When this image is read and processed, black pixels are detected only at the third raster and the second pixel. At this time, all around the target pixel (pixel A) are white ("0") pixels, and there are four vectors. That is, there are two in the horizontal direction and two in the vertical direction. In the embodiment, each time a new rough contour vector is generated, an item number is assigned to the vector. In this case, the item numbers of 0 and 1 are assigned in both the horizontal and vertical directions. Therefore, the horizontal / vertical vector table is in the state shown in FIG. The way to read this table is as follows.

Now, for example, attention is paid to a horizontal vector having an item number of “0”. The vertical vector of the item number “1” flows into the starting point of this attention vector (upper left corner of the pixel A),
This indicates that the vertical vector of the item number “0” has flowed out from the noticed end point. The item number of each vector may be considered as an offset address indicating a position stored in the table. That is, it can be seen that the data of the vertical vector flowing into and out of the horizontal vector of the item number “0” exists at offsets “0” and “1” of the vertical vector table.

Returning to FIG. 23, the pixel next to pixel A goes to the right, but since the pixel of interest is white, the process proceeds to the next pixel B without doing anything.

When attention is paid to this pixel B, it can be seen that vectors exist in the vertical and left directions except for the right side. Moreover, the upper left corner of the pixel of interest is the starting point of the horizontal vector,
It can be seen that the coordinate position of the lower left corner can be treated as the starting point of the vertical vector. Therefore, these points can be newly registered in the horizontal / vertical vector table. However, since “1” is already used as the item number in both the vertical and horizontal directions, the coordinates (3.5, 2.5) of the upper left corner of the target pixel are set as the starting point of the horizontal vector of the item number “2”. The coordinates (3.5, 3.5) of the lower left corner are stored as the starting point of the vertical vector of item number "2". If attention is paid to the horizontal vector of the item number “2”, the vector flowing into itself is the vertical vector of the item number “2”, and the flowing out vertical vector is not known at this time (the pixel following the horizontal direction). Are marked) so that they can be distinguished from each other. The same applies to the vertical vector of the item vector “2”. Thus, each table is updated as shown in FIG.

When the process proceeds to the pixel C, the coordinate positions (5.5, 3.5) and (5.
5, 2.5) occur. Therefore, the table is updated with the coordinate position. At this time, the vertical vector which has not yet been determined and flows out of the horizontal vector of the item number “2” is the vertical vector of the item number “3” newly registered this time. Write. Similarly, the vertical vector is updated. Thus, each table is updated as shown in FIG. Then, when the horizontal and vertical vectors of all the coarse contour vectors and their connection relationship are extracted,
A vector table as shown in FIG. 9 is created based on the vector tables.

FIG. 21 shows an example of an input image, and FIG. 22 shows a vector table obtained from the input image.

<Smoothing / Coordinate Conversion Circuit> Next, the smoothing / coordinate conversion circuit 900 in the embodiment will be described. As shown in FIG. 10, the smoothing / coordinate conversion circuit 900 according to the embodiment includes:
First smoothing circuit 910, first smoothing conversion table 94
0, a second smoothing circuit 920 and a coordinate conversion circuit 930.

The first smoothing circuit 910 fetches the data obtained by converting the raster scanning type binary image shown in FIG. 9 into the rough outline data sequence by the outline extraction circuit 700, and traces the coordinate sequence for each vector closed area. While referring to the first smoothing conversion table based on the vector connection state, the conversion of the outline coordinate data sequence and the labeling of the corner points are performed.
The smoothed vertex coordinate data sequence and information on whether each vertex is a corner point are output. The second smoothing circuit 920 obtains a weighted average of coordinate values of a plurality of points before and after the own point based on the first smoothed vertex coordinate data string and corner point information, and outputs a contour vector coordinate data string. . In the coordinate conversion circuit 930,
If necessary, the coordinate values of the x-axis and the y-axis are exchanged, and each coordinate value is multiplied by a scaling factor. Here, in the case of the reduction processing, the data is passed directly to the coordinate conversion circuit 930 through the smoothing processing of the first smoothing circuits 910 and 920.

The above processing will be described in detail. In the first smoothing circuit 910, as shown in FIGS. 11 to 18, the pattern of up to three sides before and after the own side is referred to,
Removal and conversion of the coarse contour coordinate numerical data sequence are performed by referring to the pattern based on the length. Here, "N_" shown in FIG.
pnt ”is the total number of closed loops of the coarse contour coordinate numerical data,
The single circle is the starting point of the horizontal vector and the ending point of the vertical vector,
A triangle indicates an end point of a horizontal vector and a start point of a vertical vector, and a double circle indicates a corner point. Also, the first here
In the smoothing, noise (notch / isolated point) peculiar to the binary image
11 are included as patterns for removing. That is, as shown in FIG. 11, in removing an isolated pixel of one pixel size, all the coordinate values are deleted. This prevents the pixel from being emphasized when the image is enlarged. As shown in FIG. 12, when white holes in a black area are stored, all outline coordinate sequences are labeled with corner points, and the coordinate values are output as they are. Here, the corner point means that the coordinate position is not moved in a second smoothing process described later. In addition, although the description is changed before and after, in the embodiment, X, Y
Since the directions are rightward and downward, the leftward and upward directions are minus directions, respectively. In addition, for example, in FIG. 13 and the like, “≦ −” of “≦ -3” indicates that the pixel continues in the minus direction, and “3” indicates that the pixel continues for at least three pixels.

The meaning of FIGS. 11 and 12 is as described above, and the following briefly describes FIG. 13 and subsequent figures.

The meaning of FIG. 13 is that a coordinate position representing an end of a thin line having a length of at least three pixels, which is a line segment having a width of one pixel, is defined as a corner point.
This means that the coordinate position is not changed by the smoothing process, in other words, rounding is prevented. In other words, in the background, a line segment having a length of 3 pixels or more and having a width of 1 pixel is recognized as a part of a character or a line drawing, not "dust" of a read image. Means that. Here, the length is 3 pixels or more, but this depends on the reading resolution.
The length may be two pixels if the reading resolution is low, or four or more if the image is read at a higher resolution. This can be said for the following explanation.

The meaning of FIG. 14 means that the unevenness of one pixel in a flat state of a predetermined length (three pixels in the drawing) is deleted, that is, the rough contour vector data of the unevenness is deleted. I do. FIG. 15 means that when the unevenness exists continuously, the unevenness is flattened.

FIG. 16 shows the definition of corner points and the concept of vector smoothing. However, in the drawing, “D _i ” indicates a coarse contour vector of interest. For example, in the case of FIG. 9A, when the end point position of the coarse contour vector of interest is in the illustrated state, it means that the end point position of the coarse contour vector of interest is a corner point. Further, attention vector D _i deletes the vector D _i-1 of the immediately preceding, the same length as the previous two vectors D _i-2 taken from the starting point of the vector of interest, its end position and the vector D _The vector connecting the start point position of _i-2 is updated as _Di-2 . Further, the vector represented by the updated end point and the corner point of _Di-2 is updated as the vector of interest.

FIG. 17 shows the contents of the smoothing process of the gentle shaded portion. For example, as shown in FIG. 9A, for an edge inclined in a predetermined direction such that it is raised by one pixel, extends three or more pixels in the horizontal direction, and rises again by one pixel, the midpoint position of the target vector is set to the immediately preceding vector D. The end point of _i-1 and the start point position of the vector D _{i + 1} immediately after are set. This processing is not performed when the inclination is in the upward direction and becomes lower in the middle.
This determination can be found by multiplying the slope of the immediately preceding vector by the slope of the immediately succeeding vector and examining the sign thereof. Also, FIG.
Shows an example in which five vectors are finally reduced to three. In other words, under the conditions as shown in the figure, the vector D before the smoothing is deleted by deleting the immediately preceding and succeeding vectors.
_i-2 vector the vectors were tied and point A on the vector D _i that the start position has focused D _i-2, attention vector represented by points A and B vectors D _i, the point A The vector represented by the end point of the vector _{Di + 2} before smoothing is Di _{+ 2}
And FIG. 18 is as illustrated.

Now, in the second smoothing circuit 920, FIG.
As shown in (2), the corner points indicated by double circles are left as they are, and the weights of the coordinate values of a plurality of points before and after the own point (in the embodiment, coordinate values with the two points before and after) with respect to the other coordinate points An average is obtained, and the value is output as a vertex coordinate data string after the second smoothing. Here, a point Q _i shown in FIG. 19 indicates a vertex coordinate data sequence after the second smoothing, and a point P _i indicates a vertex coordinate data sequence after the first smoothing. In the example, Q ₁ (9/4, 2) is changed to P ₀ (1,
3), calculated by a weighted average of P ₁ (2, 2) and P ₂ (4, 1). Here, the number of significant digits used in the first smoothing and the second smoothing is arbitrarily set within a range effective for the recording resolution.

The coordinate conversion circuit 930 performs x- and y-axis coordinate conversion depending on the difference between the main scanning of the processed image and the recorded image, and multiplies each coordinate sequence by a scaling factor in accordance with the scaling process.

For example, the input image is scanned in both main scanning and sub-scanning by 2
In the case of doubling, both the x-coordinate and the y-coordinate of the second smoothed vertex coordinate data string are doubled, and the decimal point is rounded off to perform coordinate conversion.

The outline coordinate data sequence obtained in each of the circuits 910 to 940 is output to the memory B 1100 and the processing is completed.

<Binary Image Regenerating Circuit and Intermediary Circuit> In this embodiment, a case of a page memory using type in which a bit map image memory is used for one screen will be described. Of course, all methods of converting the outline coordinate data sequence into a binary raster image are effective.

The binary image regenerating circuit 1000 reads the outline coordinate data sequence stored in the memory B 1100 of FIG.
Draw an outline on the surface.

At this time, two to three consecutive line segment vectors (each in the outline coordinate sequence) can be filled at the subsequent intermediate coating circuit at a high speed and the outline is drawn with a bitmap image of one page. Control the drawing method of the contour pixels on the line segment vector of interest by referring to the line segment vector of interest, the vector immediately before the vector, and the state (direction) of the vector immediately after the vector. I will do it. The content of the control is to treat the end point on the line segment vector of interest and the points on the contour vector other than the end point separately, and not to draw them, to draw them at the pixel positions on the line vector, It is determined whether the image is shifted to a pixel position adjacent to the pixel on the line vector in the main scanning direction by one pixel or not. The actual contour pixel drawing operation is performed by EXORing the already stored value with 1 at the pixel position to be drawn.
Is executed in the form of storing the result. In any case, the contour is drawn so that the pixels can be switched on / off only by sequentially reading the pixels in the main scanning direction. In addition, the proposal about such contour drawing is as follows.
The one already proposed by the applicant of the present application shall be adopted, and the detailed description is omitted here.

In the intermediate painting circuit 400, after the contour drawing circuit has followed the drawing processing of all outline edges, the contour drawing image in the bitmap image memory (memory A300) is raster-scanned and read at an arbitrary synchronous timing, while The middle coat processing is executed in the line processing, and the middle coat processing is transferred to the bitmap image memory again. The intermediate coating process itself is processed by a circuit as shown in FIG.
Is output to OUT while being sequentially inverted in black and white in the contour line drawing pixel.
Here, LSYNC is a line synchronization signal and is input as a synchronization reset signal at the start of raster pixel input. 5
Reference numeral 10 denotes a latch which holds the OUT output value one pixel before. 5
Reference numeral 20 denotes an EXOR, which outputs an EXOR of the input data and the OUT output value of one pixel before.

Here, other than the outline edges that are not drawn, the drawing of the outline pixels on the straight line including the processed end points is also carried out by the intermediate coating circuit.

[Explanation of Other Embodiments] In the above embodiment, the character / line image and the halftone image are separated by the image area separating circuit. However, the image attribute is previously determined by the mode select button of the image scanner or the communication protocol. If it is clear that the converted image is a pseudo-halftone or a character / line image, it is not necessary to use an area table by image area separation, and it is necessary to switch whether to perform pixel density conversion by contour extraction based on image attributes. May go. Further, the pseudo halftone image has a relatively small total number of points in one closed loop of the outline coordinate data sequence and the direction of the vector changes frequently within one closed loop as compared with the character / line drawing image.
A simple image area separation function can be realized by switching whether to perform or not perform the smoothing processing based on the total number of one closed loop of the outline coordinate data sequence at the stage of performing the smoothing processing instead of 0 and the area determination table. Specifically, the processing is performed as follows.

That is, since the minimum and maximum points of the x and y coordinate points of all the vectors forming one closed loop can be obtained, the size of the rectangle circumscribing the closed loop can be determined. That is, the size of the closed loop can be indirectly extracted as a rectangular size. Therefore, from the relationship between the rectangular size and the total number of points forming the closed loop, it is possible to determine whether or not to perform smoothing, in other words, whether the closed loop is a halftone image or a character / line image. become. In this case, the relationship between the rectangle size and the closed loop may be stored as a table or may be stored as an expression.

Further, in this embodiment, since both the enlargement / reduction processing and encoding processing based on the outline coordinate data string and the conventional enlargement / reduction processing based on the binary raster image are included, the data size of the outline coordinate data string and MH , MR, and MMR encoded data are compared, and the image data having a smaller data amount is transmitted, so that the communication time can be efficiently reduced.

Further, since the outline coordinate data sequence is considerably large depending on the image, when the capacity of the memory B1100 is small, the memory may overflow when extracting the outline coordinate data sequence depending on the image. In this case, the contour extraction processing may be stopped and the processing may be switched to the coordinate conversion (magnification) processing using the binary raster image. That is, the memory B110
Since the capacity of 0 is known in advance, it is sufficient to compare the final address position of the coarse contour vector table being constructed with the final effective address position of the memory.

In addition to the scaling by the above coordinate transformation, it is possible to multiply the scaling factors for the main scanning and the sub-scan separately, and it is possible to perform the scaling independently for the x-axis and the y-axis. In this process, the input document image size is compared with the paper size for printout, and the x-axis scaling factor and the y-axis scaling factor may be calculated based on the size difference. Of course, the user (operator) may arbitrarily set according to his / her preference. In addition, these scaling factors of the x and y axes may be processed by the transmission side by negotiation, or may be processed when the reception side prints the received data.

In the smoothing process, the above embodiment removes black isolated points and notches to remove dust pixels. However, image data read at the standard resolution of the G3 standard is used. In many cases, notches and isolated points are necessary as information for character interpretation, and therefore, a pattern for notch removal and isolated pixel removal according to the resolution of an input image in an external switching mode (for example, a switch is provided on an operation panel (not shown)). Can be removed from the smoothing pattern to perform the processing. Of course, it is possible to easily switch between performing and not performing the smoothing process according to preference using the external mode key.

In the coordinate conversion circuit 900 of this embodiment, only the exchange of the x coordinate and the y coordinate and the multiplication of the magnification are performed. By performing coordinate rotation by affine transformation of the outline coordinate data sequence, it is possible to correct skew of the read document. In this case, for example, FIG.
As shown in the figure, the read original is transported in the sub-scanning direction,
Alternatively, when scanning is performed by moving the scanner in the sub-scanning direction, edge positions of the document in the main scanning direction may be sequentially detected. That is, as shown in the figure, the displacement amount lx when one document is read and the transport amount ly in the sub-scanning direction at that time are detected.
Is obtained as tan ^-1 (ly / lx).

There are various methods for detecting a change in the edge position of the document. For example, when a document is transported and its recording surface is read, the background color at the reading position (the color detected when there is no read document) is set to a color other than white. In general, the background color of the original document itself is white, so that the edge can be detected as long as the original is distinguished from the white color. In addition, a pressing plate is provided that presses the document with a small force that does not cause the document to bend from the lateral direction of the document conveyance direction and that is movable in the left-right (main scanning) direction. The degree of skewing may be detected by detecting the left and right movement amounts of.

In the skew feeding correction processing, when an original image is read and the read image is stored in or stored in the memory A300, the image is rotated in the reverse direction by the obtained skew angle (affine transformation). It should be done. Thereafter, the above-described processing may be performed. Further, a rotation process may be performed on the coordinate data obtained by performing the above-described second smoothing process. When the skew correction is performed, such a process may be automatically performed. However, the operator is notified of the fact (a message is displayed on a display unit provided on the operation panel, etc.), and the skew is corrected. It may be instructed whether or not to correct.

Further, regardless of the skew reading of the original image, the original image may be rotated by instructing an arbitrary angle from the operation panel. That is, a case is prepared in which the characters of the document to be read are already skewed.

In the binary image regenerating circuit, the case of the page memory use type in which the bit map image memory is used for one screen has been described. However, as in the case of scan line conversion (bucket sort method), the edge table and the active edge table are used. May be used to draw a ridge edge of a contour image using a line buffer of several lines.

In this embodiment, a printer is used as an image output device.
By using a video memory and a display instead of, an image can be output to a display.

As an application example of communication using the outline coordinate data sequence, a digitizer indicating an image position is separately provided.
A decorative code (shading or change of color painting pattern, change of color) may be added to the contour coordinate data closed loop in the area designated by the digitizer and transmitted. On the receiving side, it is possible to easily change the pattern of the hatching or the inside color change and the color change according to the decoration code.

Further, instead of once developing and transmitting a font as a binary raster image such as a report or a header, the font can be directly transmitted / received as font data using an outline, or the conventional outline development can be regenerated as a binary image. It is also possible to facilitate the output of the outline font by using it in common with the intermediate coating circuit.

Of course, in the present embodiment, all the operations used in the conventional resolution conversion can be realized, and arbitrary magnification processing using a magnification specifying key and processing for putting a long original on a standard cut sheet by detecting a cassette size are possible. Can be realized by the configuration of the present embodiment.

As described above, according to the present embodiment, even when an original is read while being skewed, the skew is corrected by rotating in the reverse direction, so that an erect and good image is obtained. Can be obtained.

Further, in addition to treating as an original image bit image, the contour vector is extracted and processed based on the extracted contour vector, so that the edge does not become jagged.

Also, since data based on the extracted contour vector is transmitted to the other party, and the other party reproduces and outputs an image based on the received data, the time required for communication can be reduced, and the high-definition image can be reproduced. Transmission becomes possible.

Further, since the contour vector extracted from the original image is smoothed, even if the enlargement processing is performed, the vector is enlarged and the image is reproduced, so that a high-quality image can be obtained regardless of the magnification. Becomes possible.

Further, only when the resolution conversion is necessary, the scaling processing of the extracted contour vector is performed, so that the load on the system can be reduced as much as possible, and the normal processing speed without performing the scaling processing decreases. Can be prevented.

When extracting a contour vector from the read binary image, it is determined whether or not the storage size of the contour vector exceeds the memory capacity. If it is determined that the memory is over, the contour vector extraction processing is immediately performed. Is stopped, and the process is switched to the process using the binary image. This makes it possible to print or transmit correctly even when the memory is over.

When performing scaling processing based on contour vector data, whether or not to perform smoothing processing is switched by an external switch. This makes it possible to switch between performing smoothing processing and emphasizing processing speed according to the operator's preference. Furthermore,
When the read or received image is printed at a variable magnification, a high-quality image with little distortion can be obtained by performing the smoothing process. Moreover, if the magnification is designated by a switch (not shown), an image of a desired size can be obtained. Also, since the scaling process based on the contour vector data is performed only when performing the enlargement process, the load on the device is reduced, and it is possible to prevent the processing speed when performing processes such as the same size reduction and the like from being reduced. Will be possible.

Further, it is possible to determine to some extent whether the image is a halftone image or a character / line drawing image based on the total score of the extracted closed loop, so that it is possible to omit devices and circuits related to area determination, and to reduce cost. It will be possible to measure down.

The scaling process based on the contour vector is as follows.
This is effective only for a character / line image, and a good result cannot be obtained when it is applied to a halftone image. In this regard, according to the above-described embodiment, the scaling process is performed based on the outline vector data obtained only for the character / line image, and the scaling for the halftone image is performed by interpolating the pixels forming the image. And a thinning process, it is possible to obtain a good image.

In the communication between the devices described in the embodiments, the data to be communicated can be performed using contour coordinate data, and the transmission side or the reception side can change the magnification of the data in the x and y directions. It can be performed independently, and even if such zooming is performed, it is possible to obtain a good image according to the user's preference.

Note that the present invention may be applied to a system composed of a plurality of devices or an apparatus composed of one device. Needless to say, the present invention can be applied to a case where the present invention is achieved by supplying a program to a system or an apparatus.

[0117]

As described above, according to the apparatus or method of the present invention, when a series of processing is performed by extracting an edge vector of an image, a halftone is determined based on a vector obtained in the process. It is possible to distinguish between a binary image and a binary image.

Further, according to the apparatus and method of the present invention, it is possible to distinguish between a binary image and a halftone image with a simple configuration and reproduce a good image.

[Brief description of the drawings]

FIG. 1 is a block diagram of a facsimile apparatus of the present embodiment.

FIG. 2A is a flowchart showing the contents of a transmission process.

FIG. 2B is a flowchart showing the contents of a transmission process.

FIG. 2C is a flowchart showing the contents of a transmission process.

FIG. 3A is a flowchart showing the contents of a reception process.

FIG. 3B is a flowchart showing the contents of a reception process.

FIG. 4 is a flowchart showing the contents of a copy process.

FIG. 5 is a block diagram of a contour extraction circuit in the embodiment.

FIG. 6 is a diagram for explaining raster image scanning in the contour extraction circuit.

FIG. 7 is a diagram for explaining a horizontal coarse contour vector table.

FIG. 8 is a view for explaining a vertical coarse contour vector table.

FIG. 9 is a table of a contour coordinate data string.

FIG. 10 is a block diagram of a smoothing / coordinate conversion circuit according to the embodiment.

FIG. 11 is a diagram for describing a first smoothing process in the embodiment.

FIG. 12 is a diagram for describing a first smoothing process in the embodiment.

FIG. 13 is a diagram for describing a first smoothing process in the embodiment.

FIG. 14 is a diagram for describing a first smoothing process in the embodiment.

FIG. 15 is a diagram for describing a first smoothing process in the embodiment.

FIG. 16 is a diagram for describing a first smoothing process in the embodiment.

FIG. 17 is a diagram for describing a first smoothing process in the embodiment.

FIG. 18 is a diagram for describing a first smoothing process in the embodiment.

FIG. 19 is a diagram for describing a second smoothing process in the embodiment.

FIG. 20 is a block diagram of an intermediate coating circuit.

FIG. 21 is a diagram illustrating an example of an input image.

FIG. 22 is a diagram illustrating an example of a contour coordinate data sequence for the image of FIG. 21;

FIG. 23 is a diagram illustrating a specific image example for describing the processing content of coarse contour vector extraction in the embodiment.

FIG. 24 is a diagram for explaining a process of constructing a coarse contour vector table.

FIG. 25 is a diagram for explaining the principle of detecting the degree of skew of a document.

[Explanation of symbols]

 REFERENCE SIGNS LIST 100 Image scanner 200 Image area separation circuit 300 Memory A 400 Filling circuit 500 FIFO 600 Printer 700 Contour extraction circuit 800 Code decoding circuit 900 Smoothing / coordinate conversion circuit 1000 Binary image regeneration circuit 1100 Memory B 1200 CPU 1300 CCU 1400 Bus B 1500 Bus A 1600 Raster image scaling circuit

Claims (4)

(57) [Claims] 1. A contour extracting means for extracting contour vector data along an edge of an input image; a counting means for counting points representing vectors for each contour closed loop extracted by the contour extracting means; Detecting means for detecting an individual closed loop size based on the contour vector data obtained, and each closed loop is represented by each closed loop based on the size of each closed loop detected by the detecting means and the number of vector points corresponding to each closed loop. An image processing apparatus comprising: a determination unit configured to determine whether an image area is a halftone image or a character / line image. 2. A contour extraction step for extracting contour vector data along an edge of an input image, a counting step for counting points representing a vector for each extracted closed contour loop, Based on the size of each closed loop detected in the detection process and the number of vector points corresponding to each of the closed loops, the image area represented by each closed loop is halftoned. A determining step of determining whether the image is a character or a line drawing image. 3. An outline extracting means for extracting outline vector data along an edge of an input image, a smoothing means for smoothing the extracted outline vector data, and an outline vector data smoothed by the smoothing means. Generating means for generating an output image in accordance with the following, further comprising: a counting means for counting points representing a vector for each contour closed loop extracted by the contour extracting means; and an extracted contour vector. Detecting means for detecting the size of each closed loop based on data; and the smoothing processing means for each closed loop based on the size of each closed loop detected by the detecting means and the number of vectors corresponding to each closed loop. Control means for controlling whether or not to energize the image processing apparatus. 4. A contour extraction step for extracting contour vector data along an edge of an input image, a smoothing step for smoothing the extracted contour vector data, and contour vector data smoothed by the smoothing step And a generating step of generating an output image according to the following: a counting step of counting points representing vectors for each contour closed loop extracted in the contour extracting step; and an extracted contour vector. A detection step of detecting an individual closed loop size based on the data; and a smoothing processing step for each of the closed loops based on a size of each closed loop detected in the detection step and a vector score corresponding to each of the closed loops. And a control step of controlling whether to energize the image.