JP2796766B2 - High Efficiency Coding Method of Binary Image Mixed Natural Image - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to relates to a high efficiency coding how-described binary image mix still natural image monochrome grayscale or color images.

[0002]

2. Description of the Related Art Conventionally, various coding systems for color still images have been proposed.
Two international standard schemes called JPEG scheme and JBIG scheme, which are being studied jointly by TT SG VIII and IEC / JTC1 / SC29, are attracting attention. JPEG (Join) proposed as a color still image coding method
t Photographic Expert Group) is not an information storage type, but is a method that can encode grayscale images with high efficiency. The JPEG scheme is generally one of orthogonal transformation schemes that is an efficient scheme for a color still image, and is a discrete cosine transform (DCT).
Has been adopted. DCT is a method in which an image is divided into blocks, discrete cosine transform is performed for each block, and output coefficients are encoded. The output coefficient is quantized independently for each frequency component, and the number of quantization bits is assigned to each transform coefficient so that the coding efficiency is optimized. Unlike encoding a binary image,
In the case of a natural image, it is not practical to use an encoder having no information loss due to the amount of information included in the image, and quantization at an appropriate level is required. In general, high coding efficiency can be achieved by coarsening the quantization in the high frequency region. On the other hand, as an encoding method for a binary image, J is proposed as an international standard.
The BIG (Joint Bi-level Image Expert Group) method is an information preserving coding method capable of efficiently coding an image basically expressed in binary without loss of information.

[0003]

In a color image actually used in an office or the like, it is general that characters are mixed in the image. However, an encoding method that employs an orthogonal transformation method such as the JPEG method has a drawback that image quality of a character area is greatly deteriorated for a document in which characters and images are mixed. In Japanese Patent Application No. 3-353952, "Encoding method for binary image-mixed still natural image", an encoding method for solving the above problem is proposed. By combining the conversion methods,
It is shown that high encoding efficiency can be obtained for an image in which characters and gradation images are mixed, particularly for an image in which characters are artificially combined with a color image. Furthermore, if this method is applied to an image in which the edge area is dull or edge-enhanced by scanning with a scanner, the characteristics will differ greatly depending on the image. There is a drawback that it becomes worse. Images that require high-efficiency encoding for transmission or other purposes are not only pure natural images, but also images that contain mixed images and related printed characters, and multi-color documents. It is expected that there will be many image transmissions. For this image, the case of applying the 2 negative image mixed still image coding system in the prior art, images listed above, in particular, against the artificial image combining character color image, high code Conversion efficiency is obtained.
However, when the edge region of the image is very dull, there is a disadvantage that the number of coding bits increases.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problems of the prior art, and is intended to provide a black-and-white gradation image, a color image or a multi-color image and a natural color image including a region having a sharp edge like a character. Binary image mixed static like mixed document
For a still image, a binary image mixed natural image that can obtain an encoded image with little deterioration and can improve the encoding efficiency even for an image whose edge region is extremely dull. it is intended to provide high-efficiency encoding how the images.

[0005]

In order to achieve this object, a high-efficiency coding method for a binary natural image-mixed still natural image according to the present invention comprises a black-and-white gradation image or a color image as an original image to be coded. Is generated using a binary image mixed still image coding method, and the original image and the difference image are compared on a pixel block basis, and the binary image mixed still image coding method is different from the natural image coding method. It is configured so that only the pixel blocks evaluated as having high coding efficiency are encoded by the binary image mixed still image encoding method, and the other pixel blocks are encoded by the natural image encoding method. I have. On the encoding side of the binary image mixed still image coding method, the target original image is divided into pixel blocks, and a pixel block to which the binary image mixed still image coding method is applied and a natural image coding method are used. The block to be applied is selected in the manner described below. First, after the original image is determined to be in the character area, the binary image mixed still image coding method is applied to all blocks determined to be the character area, and the image is binarized and the gray image is created. Then, a difference image is generated. The following describes how the image quality can be improved by sending the pixel blocks of the original image and the difference image to the coding evaluation part, and applying the binary image mixed still image coding method for each pixel block. To evaluate. A quantization table for a DCT natural image according to the related art generally has a larger step size as a high frequency component. Like the edge of a character,
In a block where the pixel value changes steeply, many high-frequency components are generated, and if the quantization table as described above is used, the image quality is greatly deteriorated.

Therefore, as one method, if the change in the pixel value in the pixel block of the difference image is small as compared with the original image, the image quality is improved by applying the binary image mixed still image coding method. It is possible to use a method of estimating that it has been performed. As another method, an original image and a difference image are each subjected to DCT encoding / decoding, and before and after encoding are compared to check whether deterioration of image quality has been suppressed, thereby evaluating that image quality has been improved. Can be used. In the pixel block for which it is determined that the binary image mixed still image coding method should not be applied based on the result of the coding evaluation part, the original image is directly used as the natural image code without using the generated difference image. Become Finally,
In each pixel block, a binary image mixed still image coding method or a natural image coding method is applied. The decoding uses the same one as the decoding side of the binary image mixed still natural image coding method according to the prior art.

[0007]

Embodiment 1 FIG. 1 is carrying out the coding how according to the invention
It is a configuration diagram of a device for. 1 is an image blocking circuit,
2 is an image binarization circuit, 3 is an image synthesis circuit, and 4 is a difference image creation circuit. 5 is a binary image encoder, 6 is an encoder, 7 is a conventional natural image encoder, and 8 is a signal output circuit. Reference numeral 9 denotes a character area determination circuit, and reference numeral 10 denotes an encoding evaluation circuit. The image blocking circuit 1 converts the original image S as shown in FIG. 3 into, for example, 8 vertical pixels × 8 horizontal pixels = 64
It is divided into pixel blocks composed of pixels. The image S ′ composed of all divided pixel blocks is sequentially output to the character area determination circuit 9. The character area discriminating circuit 9 analyzes the image S ′ received from the image blocking circuit 1 for each pixel block, and performs a binarization process in the image binarizing circuit 2 only on the pixel blocks determined to include characters. Is given to the image binarization circuit 2. The image binarization circuit 2 receives the pixel blocks divided by the image blocking circuit 1, and calculates, for each pixel block, an average value Sa of the pixel values of the pixels included in the pixel block. Next, among the pixels of the image S ′, an image B is generated by converting a pixel having a pixel value larger than Sa into 1 and a pixel having a lower pixel value into 0 for each block, and the image synthesizing circuit 3 and the binary image code. To the generator 5. At the same time, for each block, the average value Sh of the pixel value of the pixel value larger than Sa and the average value S1 of the pixel value of the pixel value smaller than Sa are calculated, and the image synthesizing circuit 3 and the difference value encoder 6 are calculated. Transfer to

The image synthesizing circuit 3 compares the image B received from the image binarizing circuit 2 with the average values Sh and S for each block.
Using l, if the pixel value is 1 for all the pixels of the image B, the pixel value is replaced with the average value Sh of the block including the pixel, while if the pixel value is 0, the pixel value is replaced with the average value Sh. An image C is generated in which the pixel value is replaced with the average value Sl of the block including the pixel. The image C is transferred to the difference image creation circuit 4. The difference image creation circuit 4 includes a block image S ′ and an image C received from the image synthesis circuit 3.
, A difference image D having a pixel value obtained by subtracting the pixel value of the image C from the pixel value of the image S is transferred to the natural image encoder 7. The coding evaluation circuit 10 uses the difference image D received from the difference image creation circuit 4 and the original image S ′ received from the image blocking circuit 1 to determine the effect of the binary image mixed still image coding method. Evaluate each block. Evaluation of the results, if the effect of the binary image mixed still image coding method in the pixel block is determined to not be seen, rather than the difference image D to natural image coder 7 original image
An instruction to encode the pixel block on S ′ is given, and 2
The value image encoder 5 and the encoder 6 are instructed not to encode the image B and the average values Sh and Sl generated from the pixel block.

As an example 1 of the coding evaluation circuit 10 , the following circuit can be used. For each pixel block of the original image S ' and the difference image D , the square of the difference between one pixel in the pixel block and its vertically and horizontally adjacent pixels is calculated, and a value d obtained by adding all of them is calculated. The calculated value DS in the original image S ′ and the calculated value dD in the difference image D
Is determined by using the parameter f as follows. If dS ≧ f × dD, binary image mixed coding system dS <fxdD , natural image coding system The above-described coding evaluation circuit 10 is merely an example, and other errors caused by DCT Can be used as a circuit for comparing the following. That is, the original image
For each pixel block of S ′ and the difference image D , a pixel block that has been encoded / decoded using the conventional DCT method is generated, and the square of the change in pixel value caused by the DCT method is calculated. Calculate a value r obtained by adding all the values. The calculated value rS of the original image S ' is compared with the calculated value rD of the difference image D , and the following determination is made using the parameter f. If rS ≧ f × rD, a binary image mixed encoding method, if rS <fxrD, a natural image encoding method

[0010] The binary image encoder 5 includes an image binarization circuit 2
The received image B is encoded, and the encoded data B
c is transferred to the signal output circuit 8. As an example of the binary image encoder 5, an existing encoder such as JBIG or MMR can be used. The encoder 6 encodes the average values Sh and Sl for each block of the image received from the image binarization circuit 2 and encodes the encoded data Shc and Sl.
c is transferred to the signal output circuit 8. As an example of the encoder 6, a DPCM method that encodes a difference from an average value of a block before a continuous block can be used. Natural image encoder 7
Is the image D or the image block received from the difference image creation circuit 4.
Blocked original image that is the output of the locking circuit 1
S ′ is encoded, and the encoded data Dc is transferred to the signal output circuit 8. As an example of the natural image encoder 7 , the above-described DCT of the related art can be used. The signal output circuit 8 outputs the data Bc, Shc, Slc, encoded by the binary image encoder 5, the encoder 6, and the natural image encoder 7.
Dc and outputs the signal in a specific format.
According to the present invention, when the original image S is an image in which characters are mixed, the image S is first applied by adding the coding evaluation circuit 10.
It is possible to obtain a higher coding efficiency compared to coding using a coding method of binary mixed still natural image proposed by Japanese Patent Application 3-353952 is bright.

[0011]

[Embodiment 2] FIG. 2 is a block diagram of the embodiment 1 from which a character area discriminating circuit 9 is omitted. 1 to 8 and 10
According to the first embodiment. In the embodiment of FIG.
The pixel block is converted to a character area determination cycle as in the embodiment of FIG.
Road 9 to determine whether the block is a pixel block containing text
All of them are transferred to the binary image encoder 5 and encoded.
The individual pixel block is determined only by the evaluation judgment by the evaluation circuit 10.
Whether to use a binary image mixed encoding method for lock encoding
It is decided whether to use the image coding method. The embodiment of FIG. 2 is slightly inferior in coding efficiency to the embodiment of FIG. 1, but has an effect of obtaining an encoded image with little deterioration even for an image whose edge region is extremely dull.

The effects of the present invention The S / N ratio and the number of coded bits are reduced as shown in the following table.

[0013]

[Table 1] Here, JPEG is JPEG system, OLD is Japanese Patent Application Hei 3-35.
1 of binary mixed still image coding method proposed in 3952
For example, NEW1 is NEW2 when the evaluation circuit of Example 1 is used.
Is a case where the evaluation circuit of Example 2 is used. The definition of S / N used in Table 1 is as follows.

[0014]

(Equation 1) Here, s _{i, j} is the pixel value of the original image, d _{i, j} is the pixel value of the encoded / decoded image, and M and N are the number of vertical and horizontal pixels of the image. 255 is the maximum value that can be taken as a pixel value.

[0015]

As described in detail above, the present invention generates a differential image using a binary mixed still image encoding method for each block of an image from a target black-and-white grayscale image or color image, The original image and the difference image are compared to evaluate the image quality improvement by the binary mixed still image coding method, and only the pixel blocks having the image quality improving effect are coded by the binary mixed still image coding method. By encoding the pixel block with the natural image encoding method, it is possible to realize encoding with less deterioration of the black and white gradation image and the color image including the character area.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration example of an encoding-side device to which the present invention has been applied.

FIG. 2 is a block diagram illustrating a configuration example of another encoding side device to which the present invention has been applied.

FIG. 3 is a diagram showing an example of an original image used in the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Image blocking circuit 2 Image binarization circuit 3 Image synthesis circuit 4 Difference image creation circuit 5 Binary image encoder 6 Encoder 7 Natural image encoder 8 Signal output circuit 9 Character area discriminating circuit 10 Encoding evaluation circuit

──────────────────────────────────────────────────続 き Continued on the front page (58) Field surveyed (Int.Cl. ⁶ , DB name) H04N 1/41-1/419 H04N 7/24-7/68

Claims (2)

(57) [Claims] 1. An original still natural image is divided into a plurality of pixel blocks (n × m pixels), and a binary image area or an image block including a binary image area is detected from the plurality of pixel blocks. A binarized output is created by binarizing the pixel values of each pixel of the block using a certain threshold (Sa), as described above for each pixel block of the original still natural image.
Valued first average value of each pixel value in said each pixel block according to whether the pixel value before is small or is greater than the threshold (Sa) being the respective binarized pixel values (Sh) or a calculated value output replaced with a specific calculated value corresponding to the second average value (Sl) of the respective pixel values in each of the pixel blocks, and 2 for each pixel
Generates a difference image by using the difference between the pixel value of the before they are binarized with the calculated value output, outputs the original as it encodes and outputs a natural image coded output the difference image pictures by coding A pixel block in which the coding efficiency of a binary image mixed still image coded output is compared with the coding efficiency of each pixel block, and the binary image mixed still image coded output is evaluated to have higher coding efficiency than the natural image coded output. Only the binary image is encoded by a binary image mixed still image encoding method for encoding the binary output, the first and second average values and the difference image, and the other pixel blocks are the natural image. A high-efficiency encoding method for a binary natural image-mixed still natural image encoded by a natural image encoding system for encoding an encoded output. 2. An original still natural image is divided into a plurality of pixel blocks (n × m pixels), and a pixel value of each pixel of the plurality of pixel blocks is set to a certain threshold value (S
a) creating a binarized output by binarizing using a), as described above, for each pixel block of the original still natural image;
Valued first average value of each pixel value in said each pixel block according to whether the pixel value before is small or is greater than the threshold (Sa) being the respective binarized pixel values (Sh) or a calculated value output replaced with a specific calculated value corresponding to the second average value (Sl) of the respective pixel values in each of the pixel blocks, and 2 for each pixel
Generates a difference image by using the difference between the pixel value of the before they are binarized with the calculated value output, outputs the original as it encodes and outputs a natural image coded output the difference image pictures by coding A pixel block in which the coding efficiency of a binary image mixed still image coded output is compared with the coding efficiency of each pixel block, and the binary image mixed still image coded output is evaluated to have higher coding efficiency than the natural image coded output. Only the binary image is encoded by a binary image mixed still image encoding method for encoding the binary output, the first and second average values and the difference image, and the other pixel blocks are the natural image. A high-efficiency encoding method for a binary natural image-mixed still natural image encoded by a natural image encoding system for encoding an encoded output.