JP2546879B2 - Encoder - Google Patents

Description

The present invention relates to an encoding circuit of an electronic filing device equipped with an optical disc or the like, and particularly provides a high compression rate in accumulating drawings mainly composed of line drawings. The present invention relates to an encoding device suitable for.

[Prior Art] In recent years, an electronic filing device has attracted attention as a means for promoting paperless offices and design centers, and devices capable of accumulating larger drawings with high precision have begun to appear. In accordance with such trends, various encoding methods for determining the number of image data stored on the optical disc have been proposed. The encoding method is a distortion-free encoding method that can reproduce the original image data as it is including noise,
If the necessary information contents are retained, it can be roughly divided into a distorted coding method that does not aim to reproduce exactly as it is. 2. Description of the Related Art Conventionally, a distortion-free coding method has been widely used in a facsimile apparatus that targets an image mainly composed of characters, and a distortion coding method has been frequently used in a drawing recognition apparatus that targets an image mainly composed of line graphics.

On the other hand, a conventional electronic filing device is disclosed in
As described in Japanese Patent Laid-Open No. 59-34767, there is known one using a distortion-free coding system similar to a facsimile machine.

That is, the distortion-free encoding method for a binary image is roughly classified into a one-dimensional encoding method that encodes only in the scanning line direction and a two-dimensional encoding method that also uses the vertical direction. can do. The former is MH (Modified Huffman)
The latter corresponds to the MR (Modified READ) coding method or the MMR (Modified Modified READ) coding method.

In the MH coding method, paying attention to white or black run lengths, short codes are assigned to those having a high appearance frequency, and long codes are assigned to those having a low appearance frequency. It is composed of 0-63 bit run length or a code called terminating code,
A run length of 64 to 1728 bits is composed of a code called a makeup code.

The MR coding system identifies patterns appearing on the screen into three types of modes, that is, pass / vertical / horizontal, and assigns a variable length code to each of them. Normal image is ± 3
High efficiency coding is achieved because the rate of vertical mode coding within pixels is very high. In addition, coding is performed using the MH coding method at a rate of one for several lines. The MMR coding method is an extension of the MR coding method, and is a coding method that does not include a line using the MH coding method.

As described above, the device using the first conventional technique faithfully assigns white or black run lengths to variable-length codes, so that even complicated images such as characters can be reproduced exactly as they are. .

Further, as a conventional device for recognizing handwritten drawings, there is known a device using a distortion coding method as a feature extracting means of image data, as described in Japanese Patent Application Laid-Open No. 61-147382.

The distorted coding system cannot completely reproduce the original image,
It is expressed with a small amount of data so that the difference from the original image is hardly felt. A number of methods have been proposed, such as a method of vectorizing a line graphic image, a method of developing an image using an orthogonal exchange function, a method of dividing an image into small blocks and expressing the basic statistical distribution. Here, reference will be made to a vectorization technique that exhibits the most remarkable effect of the present invention.

In other words, the connected figure forming the input line figure is separated into a character figure and a line figure according to the features such as the size of the circumscribing rectangle and the length of the vertical line segment. Next, a thinning process is performed to convert it into a line figure having a thickness of 1 dot. Finally, the end points and bending points of the line figure are sequentially recognized, and the corresponding feature points are vector-approximated.

As described above, since the apparatus using the second conventional technique approximates the feature points of the line graphic image by vector approximation, when the line graphic image such as a drawing is targeted, the device according to the first conventional technique is used. In comparison, the information compression rate of image data can be dramatically improved.

[Problems to be Solved by the Invention] Since the prior art described in Japanese Patent Laid-Open No. 59-34767 is focused on faithfully reproducing complicated images such as characters, information of image data 2 for compression rate
Only secondary consideration was given, and there was a problem that only an information compression ratio of about 5 to 20 times could be achieved. Moreover, with the increase in size and definition of drawings, the number of stored discs per optical disc drastically decreases, which is a serious obstacle to practical use.

Since the prior art described in the above-mentioned JP-A-61-147382 focuses on faithfully vectorizing a line figure in drawing recognition, consideration is given to encoding for a complicated figure such as a character. However, although a high compression rate can be achieved for an image composed only of line figures, there is a problem that it is impossible to encode a drawing in which characters and the like are mixed.

An object of the present invention is to solve such a problem and to provide an encoding device which highly efficiently compresses a drawing in which characters are mixed.

[Means for Solving the Problems] In order to achieve the above object, the present invention is a binary image coding apparatus, wherein a distortion-free coding means for coding the binary image as it is, and the binary value. Distorted coding means for compressing only necessary information of the image and complexity detection means for determining the complexity of the binary image are provided, and the distortionless code is obtained according to the detection result of the complexity detection means. The encoding means is selectively switched between the encoding means and the distorted encoding means.

According to another aspect of the present invention, in a binary image encoding device, a plurality of types of encoding means for encoding the binary image and a complexity detection means for detecting the complexity of the binary image. Is provided, and the encoding means having the maximum compression rate is switched and selected from the plurality of encoding means according to the detection result, and the encoding is performed. is there.

As the distortion-free coding means, for example, the MH coding method, the MR coding method, or the MMR coding method, which is an international standard, is used.

As the above distortion coding means, for example, a vector coding method of a line figure is used.

The complexity detecting means detects the complexity by counting the number of times the image changes in each line in at least one axis direction for each unit block of the binarized image.

[Operation] The basic principle of the present invention is to detect the complexity of image data,
As a result of this detection, a distortion-free coding method with low compression rate but high reproduction fidelity, such as MH, MR, or MMR coding method, is applied to complex images, and reproduction fidelity is applied to simple images. At the expense of applying a distorted coding scheme with a high compression rate, such as a vector coding scheme.

The complexity is detected by counting the number of pixel changes for each unit block of the image and comparing this count value with a predetermined threshold value. If the count value is larger than the threshold value, the complexity is determined to be high, and if the count value is smaller than the threshold value, the complexity is determined to be low. This threshold may be set to an appropriate value depending on the balance between the fidelity and the compression rate. From the viewpoint of maintaining the original image, a low value is set.

According to the present invention, if the image has low complexity as a whole,
The compression rate can be dramatically increased as compared with the conventional one, and the compression efficiency can be improved even on average. In particular, in the electronic filing apparatus, most of the stored documents are drawings, so that the practical effect is remarkable.

[Example] Hereinafter, an example of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram of an embodiment according to the present invention.
1 is a scanner unit for inputting image data, 2 is an image memory unit for temporarily storing image data, 3 is a complexity detecting unit for measuring the complexity of each partial region, and 4 is only a corresponding partial region of the image data. Block division unit for temporarily storing
Is a selector for selecting inputs to both subsequent encoding units, 6 is an MH / MR encoding unit for encoding a complicated image such as a character, and 7 is a beltl unit for encoding a simple image such as a line drawing , 8 is a selector for selecting the outputs of both encoding units, 9 is an optical disk for storing compressed code data, and 10 is a control unit for controlling the entire apparatus.

In the device shown in the figure, drawings etc. written on the paper are
It is input to the scanner unit 1 and converted into binary image data by the A / D converter. The converted image data has an enormous capacity and cannot be pipelined, so it is temporarily stored in the image memory 2. By the way, the standard is 2MB when A3 drawing is sampled at 200DPI (Dots Per Inch).
It becomes the data amount of.

The image data stored in the image memory 2 is divided into sub-regions having an appropriate size and sent to the complexity detection unit 3 and the block division unit 4.

The complexity detection unit 3 scans the partial image data stored in the block division unit 4 in the X-axis direction and the Y-axis direction, and measures the number of change points. That is, the complexity of a figure is determined by measuring the frequency of change from white to black or from black to white per unit area. If the number of change points is small, it is determined that it is a line figure area, and if it is the opposite, it is determined that it is a character area. Here, the determination threshold value is set to a sufficiently small value so that all complicated figures can be included in the character area. Incidentally, in one embodiment according to the present invention, the number of change points per unit area is used as the complexity determination means, but the invention is not limited to this.

The partial image data stored in the block division unit 4 is read according to the determination result of the complexity detection unit 3, and the selector 5
Is input to the encoding unit suitable for the partial image data. That is, an image such as a line figure having a low complexity is input to the vectorization unit 7 having high compression efficiency. On the contrary, an image such as a character having a high degree of complexity is input to the MH / MR encoding unit 6 which has a low compression efficiency but good faithful reproducibility.

The MH / MR encoding unit 6 regards the partial image as an image for one page and compresses it according to the encoding procedure of the International Telegraph and Telephone Consultative Committee (CCITT). That is, in MH encoding, paying attention to white or black run lengths, short codes are assigned to those having a high appearance frequency, and long codes are assigned to those having a low appearance frequency. By using the statistical properties of image data in this way, it is possible to achieve higher compression efficiency than simple run-length coding. Further, in MR encoding, higher compression rate can be achieved by utilizing the correlation in the two-dimensional direction. The patterns appearing on the screen are identified into three types of modes, that is, pass / vertical / horizontal, and a variable length code is assigned to each of them. Since a normal image has a high correlation in the two-dimensional direction, the rate of encoding in the vertical mode having a short code length increases.
Further, in this method, in order to prevent a propagation error in an image in a two-dimensional direction, MH coding is used at a rate of one line for every several lines. The MMR coding is an extension of the MR coding and does not include a line using the MH coding, which has a large amount of information.
Therefore, in many cases, MMR coding can be realized by using MR coding hardware simply by changing the parameters of this line. Although the MH / MR / MMR coding method is used as the distortion-free coding means in this embodiment, the present invention is not limited to this.

The vectorization unit 7 vectorizes a simple line figure image and converts it into a code that does not always have good reproducibility but has only significant information that the line figure itself has. That is, thinning is applied to the connected figure forming the line figure and converted into a line figure having a thickness of 1 dot. Next, the end point of the line figure,
Recognize bending points sequentially. Finally, vector approximation is sequentially performed between corresponding end points and feature points, and the vector data is output as encoded data. In addition, as a vectorization method, there is also a method of directly vectorizing the contour line without applying thinning,
Either method may be used. In this embodiment, the vectorization method is used as the distortion coding means, but the present invention is not limited to this.

Finally, the partial image data that has been subjected to the encoding process suitable for each characteristic is stored in the optical disc 9 via the selector 8. Upon storage, auxiliary information necessary for decryption is also added and stored. That is, the size of each partial image and the encoding method are added as auxiliary information.

In FIG. 1, a control signal is supplied from the control unit 10 to each unit, and a signal notifying the state of each unit of the circuit is given to the control unit 10. However, this is not always necessary for understanding the present invention. These are not shown because they are not necessary and are not shown for simplicity.

FIG. 2 shows an operation example of the complexity detection unit 3 in FIG.
A PAD (Problem Analysis Diagram) diagram is shown. In this example, the size of one block of partial image data in the block division unit 4 is (M dots × N dots). In the figure, iCMP
Is a variable indicating complexity, iPTN (i, j) is 1 in the block
A variable that changes the value of the pixel (i, j), TH is a constant representing a threshold value, and EOR represents an exclusive OR.

In this example, first, the variable iCMP is initialized to 0 (S1),
Next, for each horizontal line in one block, the operation of comparing the values of adjacent pixels on the line is sequentially repeated from left to right. When the values of adjacent pixels are different, "1" is accumulated in the variable iCMP. (S2). A case where adjacent pixels have different values corresponds to a pixel change.

After this operation is completed, the operation of comparing the values of adjacent pixels on each line is repeated sequentially from top to bottom for each vertical line in one block.
Accumulate "1" in iCMP (S3).

The variable iCMP thus obtained represents the complexity of the block. The value of this variable iCMP is compared with an appropriate threshold value TH, and if it is equal to or greater than the threshold value TH, the MH / MR encoding unit 6 is notified of the threshold value.
If it is smaller than TH, the selector 5 is controlled so as to input the partial image data to the vectorization unit 7 (S4).

The operation of the complexity detection unit 3 is performed by software,
It can be realized by any of the hardware. Further, depending on the case, only one of the processes of S2 and S3 may be adopted.

In the image example of FIG. 3, the region 22 corresponds to one block of the image temporarily held by the block dividing unit 4 of FIG. 1, and all the images are sequentially divided into the block dividing units. Captured. By setting the appropriate thresholds for complexity detection, in this example regions 22-25 are
It is determined that the complexity is high, and the MH / MR coding method that is a distortion-free coding method is applied to each of the blocks that make up these blocks. For other areas, since the complexity is determined to be low, the vectorization method which is a distortion coding method is applied.

In other words, a simple line figure area is subjected to vectorization processing with dramatically high compression efficiency, and complicated figure and character areas are subjected to MH / MR encoding processing with poor compression efficiency but faithful reproducibility. . As is clear from this specific example, in ordinary drawings, vectorization with high compression efficiency is applied to a wide range.

From the above description, according to this embodiment, the region is divided according to the characteristics of the input image data, and the optimum encoding method is applied to each region, which is a leap compared to the conventional encoding process. The compression efficiency can be improved. This effect is particularly great in the case of drawings suitable for vectorization.

According to the present invention, the encoding circuit of the electronic filing apparatus or the facsimile apparatus is provided with the means for detecting the complexity for each partial area of the input image, and the conventional method is applied to the area showing the complicated shape. The compression rate used for MH / MR / MMR encoding, which is widely used, and the vectorization, which shows the compression rate overwhelmingly for areas with a simple shape, dramatically reduce the compression rate for drawing images. Can be improved. Particularly, in the electronic filing apparatus, the number of sheets stored on the optical disk is increased, and in the facsimile apparatus, the transmission time is shortened and its practical effect is remarkable.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment according to the present invention, and FIG.
FIG. 3 is a PAD diagram showing the operation of the complexity detector 3 of FIG.
The figure is an explanatory diagram showing an example of an input image in which an area to which vectorization is applied and an area to which MH / MR encoding is applied are separately shown. 1 ... Scanner unit, 2 ... Image memory unit, 3 ... Complexity detection unit, 4 ... Block division unit, 5 ... Selector, 6 ... MH / MR coding unit, 7 ... Vectorization unit, 8 ... Selector, 9 ... Optical disk, 10 ... Control section, 21 ... Vectorization area, 22 ... MH / MR coding area.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (56) References JP-A-60-134665 (JP, A) JP-A-60-74777 (JP, A)

Claims (3)

(57) [Claims] 1. A binary image coding apparatus, which is a distortion-free coding means for coding the binary image as it is,
Distorted coding means for compressing only necessary information of the binary image according to the vector coding method of the line figure and complexity detecting means for judging the complexity of the binary image are provided, and the complexity detecting means is provided. If the detected complexity is lower than a predetermined level, the distortion coding means is selected to perform coding, and if the complexity detected by the complexity detection means is not lower than the predetermined level, An encoding device, characterized in that the distortionless encoding means is selected to perform encoding. 2. The coding apparatus according to claim 1, wherein an HM coding method, an MR coding method, or an MMR coding method, which is an international standard, is used as the distortion-free coding means. . 3. The complexity detecting means detects the complexity by counting the number of times a pixel changes in each line in at least one axis direction for each unit block of the binarized image. 2. The encoding device according to 2.