KR20040052176A - Image compression encoding method and system - Google Patents

Abstract

PURPOSE: A method and a system for compressing and encoding an image are provided to effectively compress and encode the image including an artificial image using the substantial feature of the artificial image. CONSTITUTION: A same color pixel search unit sets an encoding object pixel in an image as a current pixel(P) according to a predetermined scan sequence(S200). The same color pixel search unit sets a search region including predetermined comparison object pixels(T) among scan pixels which exist within a uniform distance from the current pixel(P)(S205). If a pixel value(T1) of the respective comparison object pixels(T) is identical to a pixel value(P1) of the current pixel(P), the same color pixel search unit secures an increase value of a repetition count value(C) with respect to all the comparison object pixels(T)(S210). A control unit adds a corresponding increase value to a repetition count value(CO) to a previous pixel, calculates a repetition count value(CC) to the current pixel(P), and obtains the maximum value(MC) among the repetition count values(CC)(S235). If the maximum value(MC) is less than or same to the maximum value(MO) among the repetition count values(CO)(S240), the control unit encodes the maximum value(MO) and a position index value(D1) in the search region and initializes all the repetition count values(CO) as '0'(S250). If the comparison object pixel(T) in which the increase value is greater than '1' does not exist(S260), the control unit generates an encoding value(X1) of the pixel value(P1)(S270). The control unit adds the corresponding increase value to the repetition count values(CO) of the comparison object pixels(T), sets the added value as the repetition count values(CC), and generates a current pixel search result table(S280).

Description

Image compression encoding method and system {IMAGE COMPRESSION ENCODING METHOD AND SYSTEM}

The present invention relates to an image encoding method and system, and more particularly, to an image encoding method and system using characteristics of an artificial image.

Image compression coding methods for still pictures can be broadly classified into a 'lossy compression method' and a 'lossless compression method'. As a lossless compression method, the LZW (Lampel Ziv Weltch) technique is typical, and it is well known to generate a GIF (Graphic Interchange Format) file using this technique. According to the LZW technique, a pattern having a data string of a certain length is registered in a directory (code table), and when the data of the pattern registered in the directory appears, the corresponding registration number is encoded and a variable length code (Variable) is obtained. Length Code).

The compression encoding method for generating a GIF file using the LZW technique is sometimes referred to simply as the GIF method. The process of creating a GIF file involves dithering up to the number of pre-prepared color tables (e.g. 256), if necessary (i.e. when the number of colors in the image is larger than the color table) and then using the pixel values to index the color table. In this case, the variable length coding of the LZW method is performed. The GIF file generation method is a lossless compression method, and there is no change in image quality after transmission. However, in the case of natural images, the compression rate is about 30%, which is significantly lower than that of the lossy compression method (usually 90%). It is limited and is mainly used.

As a lossy compression coding method, a JPEG (Joint Picture Expert Group) method, which has been established as an international standard, is typical. The JPEG method divides image data into block units (e.g., 8x8 pixel blocks), ZigZag-Scanning (if necessary), Discrete Cosine Transform, Quantization, Huffman Coding (Huffman Coding) is a coding technique that uses a combination of techniques. Compared to the GIF method, the JPEG method is advantageous for compressing a natural image (such as a sample or a picture taken by a camera), but a lossy compression method (for example, a DCT and a quantization method). In the case of artificial images with many shapes and symbols, the image quality deteriorates in the form of damaged outlines, and the larger the quantization scale is, the more noticeable the image deterioration is. (The sample is shown in FIG. 11) Many problems appear in the compression encoding of (Artificial image).

The GIF method, also known as lossless compression, has the disadvantage of losing color during dithering when the number of colors used in the image is greater than the number of colors in the color table (typically 256 colors). Frequently, there is a need to compress and encode a mixture of an image and a natural image (shown in FIG. 12 is a sample of a mixture of natural and artificial images). It is difficult to achieve satisfactory effects using any of the compression coding methods. Some prior art attempts to overcome such problems are as follows.

Korean Patent No. 10-0205286 (Application No. 10-1995-0041666), filed on November 16, 1995 and registered on April 1, 1999, is a binary or 16 color image. Run length coding or LZW coding is performed, LZW compression coding is performed in case of 256-color CAD or graphic video, and true color coding is performed in JPEG method in case of 256-color photo image. For example, in the case of a true color image, a technique of encoding by a JPEG method is disclosed. This method has to integrate a circuit for determining the characteristics of an image to be encoded with a plurality of encoding channels, and thus the configuration of the system is complicated and the data processing takes a long time, so that there is a limitation in effectively real-time compression encoding the image data.

Japanese Patent Application Laid-Open No. 11-252563, filed on March 5, 1998 and published September 17, 1999 (Japanese Patent Application No. Hei 10-53007), is a computer graphic. A technique of dividing an image into block units for an image in which a CG) region and a natural image region are mixed, and then classifying the image characteristics of each block into four types and adaptively encoding them in four ways is disclosed. These four coding methods apply the ADCT method, the binary coding method, the binary coding method, and the block run length coding (a kind of truncation coding) which are one of JPEG methods. This method also has to be integrated with a circuit for determining the characteristics of the image to be encoded and a plurality of encoding channels, which complicates the configuration of the system and takes time to process the data, thereby limiting the effective real-time compression encoding of the image data. In addition, since coding is performed on a block basis, a problem of deterioration of an edge portion of an image may be raised.

Korean Patent Application No. 2001-0078150 (Application No. 10-2001-0004213), filed on January 30, 2001 and published on August 20, 2001, is the first mode (natural The image capturing mode) and the second mode (text image capturing mode) can be set so that the input image is encoded in the JPEG method in the first mode, and the input image (actually the color image in the second mode). The technique of compressing luminance data of data) as digitized data in the LZW method and adding essential components to the compressed data to convert the digitized image into a GIF file is disclosed.

This method simplifies the system configuration by eliminating the process of judging the characteristics of the encoding target image by adopting the user's mode selection. However, the essential characteristics of the artificial image are not found and used, but only the text data of the artificial image is processed. It is limited only in case. Therefore, this method cannot effectively encode even a simple image in which character data such as a date is recorded on a photograph of a natural image. In particular, since an actual artificial image often includes a complex image, there is a limit to image encoding in the case where a complex mixture of natural and artificial images is developed unless an effective encoding method based on the characteristics of the artificial image is developed. There is no choice but to.

As described above, a key task to overcome the limitations of the prior art is to find the essential characteristics of the artificial image and use it for effective compression coding of the artificial image portion, thereby reducing the characteristics of either natural or artificial images. It is to provide a new coding method and system capable of compressing and encoding an image effectively regardless of whether it is a strong case or a complex mixture of natural and artificial images.

SUMMARY OF THE INVENTION An object of the present invention is to provide an encoding method and system for effectively compressing and encoding an image including an artificial image by using an essential characteristic of the artificial image.

It is still another object of the present invention to provide an encoding method and a system capable of effectively performing compression encoding even when a natural image and an artificial image are complicatedly mixed regardless of characteristics of a natural image or an artificial image.

According to an aspect of the present invention to achieve the above object, an image compression encoding method comprising the steps of: setting a pixel to be encoded in an image as a 'current pixel (P)' according to a predetermined scan order; Setting a search area including a plurality of predetermined pixels among scan pixels within a predetermined distance from P in the image (hereinafter referred to as 'comparison target pixels' T)); If the pixel value T1 of each T is equal to the pixel value P1 of P, the increment ΔC of the repeat count value C corresponding to the position index D of the corresponding T is set to '1'. Otherwise, setting ΔC to '0' to secure ΔC data for all T's; Calculate 'C value up to P' (hereinafter referred to as 'CC') by adding △ C to the 'C value up to the previous pixel' (CO) of each T, Abbreviated); Generating and storing a current pixel search result table by setting CC to a value corresponding to CO of each T plus CC when the MC is greater than a 'maximum value in CO' (hereinafter abbreviated as 'MO'); If MC is less than or equal to MO, encoding and providing a 'position index value in search area' D1 of the MO and the corresponding T, and initializing all COs to '0'; After performing the initialization step, if T with ΔC equal to or greater than 1 does not exist, P1 is encoded to provide 'coded value of P1' (X1), and the corresponding ΔC is added to the CO of each T. To generate and store the current pixel search result table, and to generate a current pixel search result table by setting CC to add ΔC to CO of each T if T having ΔC equal to or greater than 1 exists. An image compression encoding method is provided.

On the other hand, according to another consistent point of the present invention, in the image compression coding system, the pixel to be encoded in the image is set to the 'current pixel P' according to a predetermined scan order, and is spatially constant from P in the image. A search area including a plurality of predetermined pixels' (hereinafter referred to as' comparison target pixels' T) among scan pixels within a distance is set, and each pixel value T1 of T is a pixel value P1 of P. ΔC data for all T's with the increment ΔC of the repeat count value C corresponding to the position index D of the corresponding T as '1', otherwise ΔC as '0'. An identical color pixel search unit to secure the same; Calculate 'C value up to P' (hereinafter referred to as 'CC') by adding △ C to the 'C value up to the previous pixel' (CO) of each T, If MC is greater than 'Maximum value in CO' (hereinafter, abbreviated as 'MO'), the current pixel search result table is generated and provided by setting the CC of each T plus △ C to CC. If MC is less than or equal to MO, it provides 'location index value in search area' (D1) of MO and corresponding T, and initializes all COs to '0', and T with ΔC greater than or equal to 1 after performing the initialization. If is not present, P1 is provided, and ΔC is added to CO of each T, and the current pixel search result table is generated by providing CC. If T with ΔC equal to or greater than 1 exists, the CO of each T is present. A controller configured to generate the current pixel search result table by setting the CC plus the corresponding? C; A pixel search result table storage unit for storing the current pixel search result table; Provided by the image compression encoding system including a basic encoding unit for generating and providing the value of the 'MO and D1' encoded as 'basic encoded data' or the 'encoded value of P1' as basic encoded data. do.

1 is a schematic structural diagram of an image compression coding system according to an embodiment of the present invention;

2 is a flowchart illustrating a development process of an image compression encoding method according to an embodiment of the present invention;

3 is a flowchart illustrating an additional image compression encoding process according to an embodiment of the present invention;

4 is a view showing a schematic configuration example of a coded data sequence generated step by step by the additional compression coding process according to an embodiment of the present invention;

5 is a diagram illustrating an example of setting a search area according to an embodiment of the present invention;

6 illustrates an example of data related to a search result table configuration according to an embodiment of the present invention;

7 is a diagram illustrating a raster scan sequence according to an embodiment of the present invention;

8 illustrates a left pixel width first scan order according to an embodiment of the present invention;

9 is a block diagram raster scan order according to an embodiment of the present invention;

10 shows an artificial image sample,

11 shows a natural image sample,

12 illustrates an image sample in which a natural image and an artificial image are mixed.

<Description of the code | symbol about the principal part of drawing>

100: basic encoding channel 102: same color pixel search unit

104: pixel search result table storage unit 105: control unit

106: basic encoder

S200: setting the encoding target pixel as the current pixel P

S205: setting a search area based on the position of P

S210: securing an increment ΔC of the 'repeat count value' C corresponding to the position index D of the 'compare target pixel' T

S230: Checking whether DELTA C is secured for all Ts in the search area

S235: calculating a repeating count value up to P (CC) by adding the corresponding ΔC to the 'repeating count value up to the previous pixel' CO of each T to obtain a maximum value MC among them

S240: comparing the MC with the 'maximum CO' (MO)

S250: a step of encoding and providing a 'position index value in a search area' (D1) of the MO and the corresponding T, and initializing all COs to '0'

S260: Checking whether T with ΔC is 1 or more exists

S270: generating and providing a 'coded value of P1' (X1) by encoding a 'current pixel value' (P1)

S280: generating and storing a 'current pixel search result table' by setting CC equal to that of CO of each T plus CC

First, the core focus of the present invention will be described. In the present invention, in order to increase the efficiency of compression coding of images, the characteristics of natural and artificial images are compared and analyzed, and the following facts are found through numerous compression coding experiments.

First, in compression coding of an image, there is a high probability that there are pixels having a color value that is the same (or similar) as the current pixel among pixels that are spatially close within the image frame. In particular, in the case of artificial images, various figures and symbols are used. Almost all shapes used by humans are represented by 'horizontal lines (-), vertical lines (|), diagonal lines (/), and historical lines (\)' in the image. For convenience, it is important to pay attention to the property of the combination of 'four basic elements of artificial image'.

With reference to the artificial image sample of FIG. 10, the advantages can be easily understood. Artificial images usually consist of lines or curves of various shapes, as well as letters and numbers, and it can be understood that if they are chopped up, they can eventually be decomposed as four basic elements of the artificial image.

Second, even if there are no pixels with the same (or similar) pixel value within a spatially constant distance, there is a high probability that the previously scanned pixel value will reappear. In particular, many of these characteristics appear in artificial images, and in the case of natural images, this principle is well applied because there are many symmetrical images.

Accordingly, the present invention pays attention to the characteristics of such an image, and the value of a predetermined pixel within a spatially constant distance having a high probability of having the same (or similar) value as that of the current pixel in the image (described in detail below). By using the result of comparing the current pixel values for encoding the current pixel values, it is possible to significantly improve the compression encoding efficiency of the image data. Hereinafter, exemplary embodiments of the present invention will be described with reference to the accompanying drawings.

1 is a schematic structural diagram of an image compression coding system according to an embodiment of the present invention. Referring to FIG. 1, an image compression encoding system according to the present invention essentially includes a basic encoding channel 100, and according to an embodiment, the first selector 110 and the queue buffer index encoder 120 may be used. The second selector 130, the lossy compression encoder 140, the pixel value difference encoder 150, the third selector 160, and the data formatter 180 may be further included. According to an exemplary embodiment of the present invention, the basic encoding channel 100 includes the same color pixel search unit 102, a pixel search result table storage unit 104, a control unit 105, and a basic encoding unit 106.

2 is a flowchart illustrating a development process of an image compression encoding method according to an exemplary embodiment of the present invention. 3 is a flowchart illustrating a development process of a compression encoding method added according to an embodiment of the present invention. 4 is a diagram schematically illustrating an example of a configuration of an encoded data string generated step by step by an image compression encoding process according to an embodiment of the present invention. Hereinafter, the development process of the image compression encoding method according to an embodiment of the present invention will be described in detail with reference to the structure and function of the compression encoding system corresponding thereto.

Referring to FIG. 2, in the development process of the compression encoding method according to an embodiment of the present invention, the encoding target pixel in the image is set to the current pixel (hereinafter, referred to as 'P') according to a predetermined scan order in step S200. To begin. The image is provided to an input terminal (not shown) of a system from a digital image input device, a digital image receiving device, a storage device, or from an external digital image data source such as a PC, and the like, conventionally, a GIF method or Although it may be referred to as a still image encoded by the JPEG method, it is not limited thereto.

The predetermined scan order may include various raster scan orders, a left pixel width priority scan order, and a block raster scan order. Here, 'raster scan order' is a term well known in the art and starts with the leftmost pixel in the top row and scans to the right as shown in FIG. 7, and starts with the leftmost pixel in the next row and scans to the right. Means the order. 8 is a diagram illustrating a 'left pixel width first scanning order', which processes all lines from the top line to the bottom line by a predetermined H pixel width (H is usually 8) from the left to the right of the image screen. Again, the H pixel width on the right is scanned in the same way. FIG. 9 is a block raster scan order in which a raster scan order is applied in HxH block units, but a pixel is scanned in a raster scan order in a block. In some embodiments, a scan order different from the above described scan order may be applied. After step S200, the process proceeds to step S205.

In step S205, the search area is set based on the spatial position of P set in step S200. Here, the 'search area' means' a plurality of scan pixels (that is, scanned pixels) within a predetermined distance (for example, a 4 pixel radius) spatially from the current pixel P in the image (ie, a scanned pixel) (usually 4-6). "Pixels" (hereinafter referred to as "comparison target pixels").

According to a preferred embodiment of the present invention, in accordance with the scanning principle of the left column first and the upper row first, as the 'search area', the left pixel of P, the upper row of P, the left pixel of P, the upper row of P It may be an area having four pixels of the right pixel of the pixel as a 'compare target pixel' (hereinafter, referred to as 'T'), which is a search area setting considering the 'four basic elements of the artificial image' described above. According to an exemplary embodiment, an area having six pixels as a total of T including an upper row pixel of the upper row of P and a left pixel of the left of P may be set as a 'search area' in addition to the four pixels.

Meanwhile, depending on the position of the current pixel, a pixel set as the search region of P but not present in the actual image or a pixel which has not yet been scanned may exist as the comparison target pixel. In this case, as the pixel value of the T, a value used in the encoding process of the previous scan pixel may be used as it is, or a specific value (for example, black or white) arbitrarily set may be used, or may be processed as 'uncomparable'. Alternatively, for convenience, it is also possible to use the immediately preceding pixel (that is, the pixel just scanned on the basis of P) as the left pixel, and the immediately preceding pixel (that is, the pixel scanned immediately before the basis of P as the left pixel). .

Step S210 after step S205 is to secure, for each T, an increment ΔC of the repeat count value C corresponding to the position index D of T (in space in the image). . More specifically, in step S210, the pixel value of each T (hereinafter referred to as 'T1') is compared with the pixel value of P (hereinafter referred to as 'P1'), and if T1 is equal to P1, the corresponding ΔC Is denoted by K (a predetermined positive integer, usually 1, so 1 will be represented below) and? C is set to '0'. Here, 'if not' means that T1 does not have the same value as P1, not only when T1 is different from P1, but also when T in step S205 is not present in the image or 'uncomparable' as described above. It should be noted that 'in any case'.

Subsequently, in step S230, if step S210 has been performed for all the Ts in the search area, the process proceeds to step S235; otherwise, the process returns to step S210. In this way,? C data for all Ts are secured. Hereinafter, the steps S205, S210, and S230 will be described in more detail with reference to FIGS. 5 and 6.

According to an exemplary embodiment of the present invention, as shown in FIG. 5, when the position of the current pixel is '0', the pixels displayed as '1' to '6' are used as the comparison target pixel. In this case, the order of searching is given priority to the smaller number. Meanwhile, according to another exemplary embodiment of the present invention, a pixel (hereinafter referred to as pixel '0') of a current pixel position of a reference image (for example, the image immediately before the current image) may be further provided as a comparison target pixel. . 5 illustrates examples of a 'search area for a previous pixel' and a 'search area for a current pixel' according to the above-described search area setting rule. In FIG. 5, the position indicated by 'N' in the search region of the previous pixel (the position of the next encoding target pixel of the current pixel position indicated by '0' in the search region of the previous pixel) is set as the current pixel position and accordingly It can be seen that the location number for the

The same-color pixel search unit 102 reads the data of the immediately preceding (scan) pixel search result table (see the example in FIG. 6A) stored in the pixel search result table storage unit 104 in advance, and based on this, T1 and P1 are compared in order according to a predetermined comparison order. According to an exemplary embodiment, the same color pixel search unit 102 may store the T1 values of all the Ts of the immediately preceding pixel in a buffer (not shown) in its own and may be used to compare P1 and T1. As can be seen in FIG. 6A, the repetition count value C is recorded in the pixel search result table corresponding to the position index D of T. FIG.

If T1 is equal to P1 (i.e., the colors are the same), the iteration count value C corresponding to the corresponding D in the search result table is increased by 1 (i.e., ΔC is 1), otherwise C is increased. (That is, ΔC is zero). This operation is continued for all the pixels of the search area in a predetermined order. In this way, DELTA C data for all Ts are secured.

A process of acquiring [Delta] C data for T of P performed by the same color pixel search unit 102 will be described in more detail with reference to FIGS. 5A and 5B. First, the pixel immediately preceding the pixel search region because the pixel (indicated by 'N' pixel) immediately to the right of pixel 0 of the search region with respect to the previous pixel becomes the pixel position 0 of the search region with respect to the current pixel. Pixels 1, 2, and 3 of the pixels correspond to pixels 6, 4, and 2 of the current pixel search region, respectively (see FIG. 5 for comparison). On the other hand, pixels 4, 5, and 6 of the immediately preceding pixel search region are located outside the search region of the current pixel and thus are not compared with the current pixel value.

Also, it can be seen that pixels 3 and 5 as the search area pixels based on the current pixel are not included in the search area of the immediately preceding pixel. After step S230, in step S235, ΔC is added to 'C value up to the previous pixel' of each T (i.e., C value in the previous pixel search result table as hereinafter referred to as 'CO'). Calculate the C value '(hereinafter referred to as' CC ') to obtain the maximum value of the CC' (hereinafter abbreviated as 'MC'). 6B shows an example of data obtained during the search for the current pixel. Referring to FIG. 6B, it can be seen that only ΔC for pixels 3 and 4 of the search region for the current pixel are set to 1, respectively. After step S235, the process proceeds to step S240.

In step S240, the step of comparing the MC with the 'maximum value of CO' (abbreviated as 'MO'). If MC> MO, the process proceeds to step S280. If MC is equal to or smaller than MO, step S250. Proceed to In step S250, a value obtained by encoding the 'position index value in the search area' D1 of the MO and the corresponding T is provided as 'basic encoded data' and all COs are initialized to '0'. For example, in the case of Fig. 6B, MO is 4 and D1 is 1. However, when two or more MOs exist, the selection is made according to a predetermined comparison priority order (for example, the small number order in FIG. 5) or the scan priority order. When all COs are initialized to 0 in FIG. 6B, data as shown in FIG. 6C is obtained. After step S250, the process proceeds to step S260.

In operation S260, as a result of the search for the current pixel, it is determined whether one or more Ts having ΔC equal to or greater than 1 exists. When one or more Ts with ΔC equal to or greater than 1 exist, the process proceeds to operation S280. Otherwise proceed to step S270. In operation S270, the current pixel value P1 is encoded, and the encoded value X1 is generated and provided as 'basic encoded data'. The process proceeds to step S280 after step S270.

In step S280, the current pixel search result table is generated and stored by setting the CC of T plus CC corresponding to CO of each T. For example, FIG. 6D illustrates an example of the current pixel search result table generated by performing step S280 based on the data of FIG. 6C. After step S280, the process goes to step S290.

In step S290, it is determined whether encoding is completed for a predetermined number (J) of P, and if it is completed, the process is terminated. If not, the process returns to step S200 and the basic for J pixels is determined. The process continues until the encoding is completed. Here, J is a multiple of 8 and preferably can be set to one of 8, 16, 32, 64. In this way, the basic encoding is completed by performing the above-described steps for each of the J pixels for the pixels in the image to generate 'J basic encoded data sequences' consisting of J basic encoded values.

The above steps S235, S240, S260, S280, and S290 may be performed by the controller 105. The controller 105 receives the current pixel search result data from the same color pixel search unit 102. After the MC is obtained in operation S235 and MC> MO in operation S240, operation S280 is performed to generate a search result table up to the current pixel and provide the result table to the pixel search result table storage unit 104 therein. To be stored. Meanwhile, the controller 105 provides the MO and D1 to the basic encoder 106 when the MC is equal to or smaller than the MO. In addition, under the condition that MC is equal to or smaller than MO, the controller 105 provides P1 to the basic encoder 106 when there is no T with ΔC equal to or greater than 1, and performs the operation of step S280 to perform the current pixel search result table. Is generated and provided to the pixel search result table storage unit 104 to be stored there. The controller 105 may control the same color pixel search unit 102 and the basic encoder 106 according to the determination of step S290.

The basic encoding unit 106 receives the 'MO and D1' or P1 from the control unit 105 and performs basic encoding on the 'MO and D1 encoded values' (hereinafter referred to as 'MOD1') or 'X1', respectively. Create and provide data. The basic encoding unit 106 normally provides a group of J encoded data obtained by performing basic encoding on the J pixels as a 'basic encoded data string'. However, in some embodiments, the basic encoding unit 106 The basic coded data may be provided in units of pixels, and the first selector 110 to be described later may configure the 'basic coded data stream' by grouping J coded values. Here, lossless compression coding is used for basic coding, and fixed run length coding (FLC) is typically used.

Meanwhile, according to a preferred embodiment of the present invention, an additional encoding process may follow after step 290. 3 is a diagram illustrating an example of an additional encoding process according to a preferred embodiment of the present invention. According to an embodiment of the present invention, in order to implement such an additional encoding process, as shown in FIG. 1 in addition to the basic encoding channel 100, the first selector 110 and the queue buffer index encoder ( 120, the second selector 130, the pixel value difference encoder 150, the third selector 160, the lossy compression encoding channel 140, and the data formatter 180 are images of the present invention. It may be further included in the compression encoding system.

4 is a schematic configuration example of an encoded data string generated step by step by an additional compression encoding process according to an embodiment of the present invention. Hereinafter, an additional process of the image compression encoding method according to the preferred embodiment of the present invention will be described in detail with reference to the structure and function of the added circuit.

First, in step S300, the data size (hereinafter, 'A' is a data column consisting of J original pixels' (hereinafter, abbreviated as' J original pixel data columns' and indicated as' 400 'in FIG. 4). (Usually expressed in bytes) and the data size of the 'basic encoded data string' (indicated by '410' in FIG. 4) provided from step S290, Select the subsequent processing path. That is, in step S300, if the value B1 / A obtained by dividing B1 by A (hereinafter referred to as 'Y1') is smaller than the first predetermined reference value TH1, the compression ratio by the basic encoding is better than the desired level. If it is determined, the process proceeds to step 310, otherwise proceeds to step S360. TH1 here is a value between 0 and 1, Preferably it is a value of 0.3-0.5 (for example 1/3).

In this way, data is compressed by basic encoding, for example, V3, V4, and V5, which are the 3rd, 4th, 5th pixel values of the data stream 400, are 24 bits (R, G, B, or Y for each pixel). , A total of 72 bits are required for the representation of the U and V values, respectively, whereas 'the data indicating that the repetition count value of the position index 4 is 3' in the data string 410 having the fixed length coding 'P4R3' is easily understood in that it can be expressed as 8 bits, for example. For reference, P4R3 decodes the decoding unit (not shown) by taking three times the reconstructed pixel value at the fourth position with respect to the pixel when reconstructing the V3, V4, and V5 values.

In step S310, the queue buffer index encoding for the 'basic encoded data sequence' is performed by the queue buffer index encoder 120 to generate a 'queue buffer index encoded data sequence'. In the queue buffer index encoder 120, there is a queue buffer (not shown) prepared separately, which is a pixel encoded by a pixel value without being encoded by a position and an iteration count in the basic encoder 106. The pixel values of pixels whose same pixel values are not stored in the queue buffer are stored in the scanning order. In more detail, step S310, a storage position index (hereinafter, abbreviated as 'queue buffer index') in a queue buffer in which a pixel equal to a current pixel value is stored is encoded and stored at a corresponding position of the queue buffer index encoded data string. . If none of the pixels stored in the queue buffer have the same pixel value as the pixel value of the current pixel, the current pixel value is inserted (stored) in the queue buffer, and an encoded value (for example, FIG. 4, X1, X2, X7, X14 of '430' are left as it is. In Fig. 4, the &quot; queue buffer index coded data stream 430 &quot; thus obtained is illustrated.

For example, referring to FIG. 4, if the value encoded by the current pixel as the 13th pixel is X13 and the pixel value is V13, the encoded value of the previous pixel is X1, X2, X6, X7, X12. For example, in the column 430, coded Q1 instead of X6 means that V6 has the same pixel value as V2. At this time, V1, V2, V7, and V12 are stored in the queue buffer. The number 1 on the right side of Q indicates the position index of the immediately preceding pixel. In this way, the storing of Q3 instead of X13 in the data string 430 indicates that the pixel value of the second pixel V2 stored in the queue buffer three times before the thirteenth pixel is the same as the pixel value of V13.

Subsequent to step S310, in step S320, a subsequent processing path is selected by comparing A with the data size of the queue buffer index encoded data stream (for example, 430) (hereinafter referred to as 'B2'). That is, in step S320, if the value B2 divided by A (B2 / A) (hereinafter referred to as 'Y2') is smaller than the predetermined second reference value TH2, the compression ratio by the queue buffer index encoding is higher than the desired level. If it is determined that it is good, the process proceeds to step 330, otherwise proceeds to step S360. TH2 is a value smaller than TH1 and is a value between 0 and 1, Preferably it is a value of 0.2-0.4 (for example, 1/4).

In operation S330, the pixel value difference encoding unit for the queue buffer index encoding data sequence is performed by the pixel value difference encoding unit 150 to generate the pixel value difference encoding data sequence. That is, the error DF between the pixel value of the pixel remaining as the pixel value still encoded in the queue buffer index coded data column and the previous pixel value is calculated, and if DF is equal to or smaller than the predetermined reference value TE, DF is calculated. Select (code fixed length) or leave the encoded pixel value as it is. 4 shows an example 450 of the &quot; pixel value difference coded data stream &quot; thus generated. In Fig. 4, those indicated by E2, E7, and E14 represent fixed-length encoded values of pixel value differences. Where E2 is the error between V2 and V1, E7 is the error between V7 and V6, and E14 is between V14 and V13. After step S330, the process proceeds to step S340.

In step S340, a subsequent processing path is selected by comparing A with the data size (hereinafter, denoted as 'B3') of the 'pixel value difference encoded data string' (for example, 450). That is, in step S340, if B3 divided by A (B3 / A) (hereinafter referred to as 'Y3') is smaller than the predetermined third reference value TH3, the compression rate due to pixel difference value encoding is higher than a desired level. If it is determined that it is good, the process proceeds to step 370, otherwise proceeds to step S360. TH3 is a value smaller than TH2 and is a value between 0 and 1, Preferably it is a value of 0.1-0.3 (for example, 1/5).

In step S360, a lossy compression encoding method is applied instead of a lossless compression encoding method such as basic encoding, queue buffer index encoding, and pixel value difference encoding. <Coded data information indicating the number of lossy compression coded pixels> (e.g., J is the number of lossy compression coded pixels) at a corresponding position between the queue buffer index coded data stream and the pixel value difference encoded data stream. In this case, inserting &quot; S64 &quot; and then storing the corresponding original pixel data stream (e.g., 400 in FIG. 4) in the lossy compression wait buffer (not shown), but waiting for the lossy compression wait buffer. When the number of pixels is greater than or equal to a predetermined number of blocks (for example, 64 pixels constituting an 8x8 block), the corresponding blocks are loss-compressed and encoded by ' The lossy compression coded data stream 'is obtained.

In this case, when the lossy compression encoding method is applied continuously, the lossy compression encoded data is continuously added as one lossless compression encoded data string until lossless compression encoding is applied. For example, when the number of pixels of the data string to be subjected to lossy compression coding is 64, and the value indicating that the number of pixels of the data string subjected to the lossy compression coding immediately before is 192 is S192, the two are successively grouped and expressed as S256.

Here, as a lossy compression coding method, a block-by-block coding method, which is a well-known JPEG method, is a method of sequentially applying discrete cosine transform (DCT), quantization, and Huffman coding, but other lossy compression coding. The method can also be used. At this time, the block size is predetermined and mainly adopts the HxH pixel size (H is usually 8 as a multiple of 8). However, if the original pixel data string that is the last lossy compression encoding target of the image is input to the lossy compression encoding channel 160, even if the HxH pixel values are not all filled, the remaining values are filled with '0' and then the block unit loss is performed. Compression coding may be performed.

On the other hand, the above-mentioned reference value (TH1, TH2, TH3) may be set differently according to the memory capacity and the calculation performance of the coding system (for example, PC). In addition, it may be controlled differently according to a user's menu selection and system setting. For example, when applying quantization in lossy compression coding according to a user's menu selection or a change in system settings, a selectivity reference value of whether the quantization scale is to be enlarged (ie, coarse encoded) or smaller (finely encoded). This can influence the decision.

Step S300 is performed by the first selector 110, and compares the size of the original pixel data string input therein with the data size of the basic encoded data string received from the basic encoder 106, and Y1 <TH1. The basic encoded data is provided to the queue buffer index encoder 120, and otherwise, a 'lossy compression encoding command' is provided to the lossy compression encoding channel 140. Step S320 is performed by the second selector 130, and compares the size of the original pixel data string input thereto with the data size of the queue buffer index encoded data string received from the queue buffer index encoding 120, and Y2. If <TH2, the queue buffer index coded data stream is provided to the pixel value difference encoder 150, otherwise, a 'lossy compression coded command' is provided to the lossy compression coded channel 140. The unit 160 performs a comparison between the size of the original pixel data string input therein and the data size of the 'pixel difference coded data string' received from the pixel value difference encoding unit 150, and if Y3 <TH3, The pixel value difference coded data stream 'is provided to the data formatting unit 180, and otherwise, a' lossy compression coding command 'is provided to the lossy compression coding channel 140.

Step S360 responds to the 'lossy compression coding command' from the first selector 110, the second comparator 120, and the third selector 160 inputted therein by the lossy compression encoding channel 140. By performing lossy compression coding on the J original pixel data streams, a 'lossy compressed coded data stream' is generated and provided to the data formatting unit 180. The lossy compression coding channel 140 includes a DCT unit for generating DCT coded data by performing DCT encoding by collecting J original pixel data streams in block units, and a quantization unit for generating quantized data by performing quantization on the DCT coded data. The Huffman encoder may perform Huffman encoding on the quantized data.

In operation S370 performed by the data formatting unit 180, a lossy compressed coded data stream and a lossless compressed coded data stream obtained as a result of encoding all of the entire image screens in units of J pixels, that is, the basic encoded data stream, The queue buffer index encoded data string and the pixel value difference encoded data string) are combined to generate 'coded image data' as encoded data strings for an entire image, thereby ending the process.

Here, according to an exemplary embodiment of the present invention, the portion of 'lossy compressed coded data values' (sequentially) collected in order (separately)' (for example, in the data stream 470 of FIG. 'WW ... WW') and 'data information for the lossy compression encoding (coded value)' (for example, a portion indicated as 'S32, S16, S64', etc. in the data string 470 of FIG. Note that it is also possible to configure this data information inserted at the corresponding position of the lossless compressed coded data string as an actual lossless compressed coded data string. In other words, the lossless compressed coded data stream should be understood as such a broad meaning.

Meanwhile, according to an exemplary embodiment of the present invention, the LZW encoding is performed on the 'lossless compressed coded data stream' immediately before data formatting, and then, in step S470, the LZW-coded data stream and the lossy compressed coded data stream are performed. It should be noted that the encoded image data may be generated in combination (described in detail in the example of generating the encoded image data 480 of FIG. 4 hereinafter).

In addition, according to the embodiment, the operation of comparing the corresponding compression rate in the first selection unit 110, the second comparison unit 120, and the third selection unit 160 by the user's menu selection or system setting. The basic encoding unit 106, the queue buffer index encoding unit 120, and the pixel value difference encoding unit 150 encode the basic encoding data sequence, the queue buffer index encoding data sequence, and the pixel value difference encoding. It should be noted that the data streams' may be provided to the data formatting unit 180 directly through the line L17 as' lossless compressed coded data streams', respectively.

In FIG. 4, '470' and '480' are illustrated as encoded image data strings for the entire image generated through data formatting in operation S370. The coded image data stream 470 is an example in which LZW is not applied, and U1, U2, U3, and U5000 are pixel value difference encoded data streams, and S32, S16, and S64 are pixels processed by a lossy compression encoding process. Loss-compressed coded data information (coded value). That is, S16 indicates that 16 pixels have been loss compression coded. In the data strings 470 and 480, the portion labeled 'WW..WW' indicates a portion in which lossy compression coded pixel values are sequentially collected.

Meanwhile, the portion denoted as 'ZZZZ' in the encoded image data string 480 applies LZW encoding to the lossless compressed encoded data string (U1, S32, U2,... U5000) in the data string 470. Shows the LZW coded data string obtained.

For reference, in the decoder (not shown), since the decoded data of the lossy compression coded data string should be used in the decoding process of the lossless compression coded data string, the decoder first decodes the lossy compressed coded data string and stores it in the buffer for a while. The efficiency of decoding can be improved by processing the pixel value required in the decoding process of the compressed coded data stream by using a method of obtaining from the lossy compression decoding buffer.

As an example of the experimental results according to the embodiment of the present invention described above, a bitmap 2.25 MB size of a 1024 x 768 RGB 24-bit computer screen image in which artificial and natural images obtained by accessing a specific Internet site is mixed. In the conventional 256-color GIF method, the data size was 156 KB. In the conventional method, JPEG compression encoding was performed with a smaller quantization scale (JPEG # 1) and JPEG compression encoding with a larger quantization scale. (JPEG # 2) showed data sizes of 232 KB and 101 KB, respectively. On the other hand, as a result of compression coding on the computer screen image according to an embodiment of the present invention, the compression ratio was the best by obtaining a data size of 71 KB, and in terms of the image quality of the restored screen, JPEG # 2 as well as JPEG # 1 It was excellent and hardly different from the conventional GIF method.

Although specific embodiments have been described in the above description of the present invention, those skilled in the art can make various modifications without departing from the scope of the present invention as defined by the following claims.

According to the present invention, an essential feature of an artificial image is used to effectively compress and encode an image including an artificial image, and whether the natural image or the artificial image is a complex mixture of natural or artificial images. Provided are a compression encoding method and a system capable of compressing and encoding an image effectively.

Claims (22)

In the image compression coding method, Setting the encoding target pixel in the image as a 'current pixel P' according to a predetermined scan order; Setting a search area including a plurality of predetermined pixels among scan pixels within a predetermined distance from P in the image (hereinafter referred to as 'comparison target pixels' T)); If the pixel value T1 of each T is equal to the pixel value P1 of P, the increment ΔC of the repeat count value C corresponding to the position index D of the corresponding T is set to '1'. Otherwise, setting ΔC to '0' to secure ΔC data for all T's; Calculate 'C value up to P' (hereinafter referred to as 'CC') by adding △ C to the 'C value up to the previous pixel' (CO) of each T, Abbreviated); Generating and storing a current pixel search result table by setting CC to a value corresponding to CO of each T plus CC when the MC is greater than a 'maximum value in CO' (hereinafter abbreviated as 'MO'); If MC is less than or equal to MO, encoding and providing a 'position index value in search area' D1 of the MO and the corresponding T, and initializing all COs to '0'; After performing the initialization step, if T with ΔC equal to or greater than 1 does not exist, P1 is encoded to provide 'coded value of P1' (X1), and the corresponding ΔC is added to the CO of each T. To generate and store the current pixel search result table, and to generate a current pixel search result table by setting CC to add ΔC to CO of each T if T having ΔC equal to or greater than 1 exists. Containing Image compression coding method. The method of claim 1, wherein all the steps are performed on the J original pixel columns to generate 'J basic coded data columns', wherein J is a predetermined multiple of 8. Image compression coding method. The method of claim 1, The search region further includes a pixel at a specific position in a previously encoded image or a predetermined reference image. Image compression coding method. The method of claim 1, The predetermined distance is a 4-pixel distance, and if the predetermined scan order follows the left column first and upper row first principles, the search area includes a left pixel of P, an upper pixel of P, a left pixel of P, and an upper row of P. The right pixel, the upper row pixel of the upper row of P, and the left pixel of the left of P are provided as the comparison target pixel. Image compression coding method. The method of claim 1, The predetermined distance is a 4-pixel distance, and if the predetermined scan order follows the left column first and upper row first principles, the search area includes a scan pixel immediately before P, an upper pixel of P, a left pixel of P, and an upper row of P. The right pixel of, the top row pixel of the upper row of P, and the immediately preceding scan pixel of P are provided as the comparison target pixel. Image compression coding method. The method of claim 1, When the pixel to be compared in the search area is a pixel that does not actually exist or a pixel that has not yet completed scanning, it is determined as 'comparable', and the corresponding ΔC is set to '0'. Image compression coding method. The method of claim 2, J is any one of 8, 16, 32, 64 Image compression coding method. The method of claim 1, The predetermined scan order may be any one of a raster scan order, a left pixel width priority scan order, and a block-by-block raster scan order. Image compression coding method. The method of claim 2, By comparing the data size of the 'J original pixel data columns' (hereinafter referred to as 'A') and the data size of the 'basic encoded data column' (hereinafter referred to as 'B1'), Y1 (= (B1 / A )) Is larger than the first predetermined reference value TH1, 'J' original pixel data strings are collected, and lossy compression encoding is performed in units of blocks having a predetermined size to generate a 'lossy compressed encoded data string'. Generating a queue buffer index encoded data string by performing queue buffer index encoding on the basic encoded data string, wherein TH1 is a value between 0 and 1 Image compression coding method. The method of claim 9, Comparing the data size of the 'cue buffer index encoded data stream' (hereinafter referred to as 'B2') and A, if Y2 (= (B2 / A)) is larger than the second predetermined reference value TH2, the 'J' Collect the original pixel data streams' and perform lossy compression encoding on a block basis of the predetermined size to generate a 'lossy compressed data stream', or otherwise perform 'pixel-value difference encoding' on the queue buffer index encoded data stream. Generating a 'pixel value difference encoded data stream', wherein TH2 is a value between 0 and 1 less than TH1; Image compression coding method. The method of claim 10, Comparing the data size of the 'pixel value difference coded data stream' (hereinafter referred to as 'B3') and A, if Y3 (= (B3 / A)) is larger than the third predetermined reference value TH3, the 'J' Collecting the original pixel data streams' and performing lossy compression encoding on a block basis of the predetermined size to generate a 'lossy compressed data stream'; otherwise, providing the pixel value difference encoded data streams, TH3 is a value between 0 and 1 less than TH2 Image compression coding method. 10. The method of claim 9, wherein the lossy compression coding on a block basis is performed by sequentially using DCT, quantization, and Huffman coding techniques. Image compression coding method The method of claim 11, Generating LZW encoded data strings by applying LZW encoding to the pixel value difference encoded data strings; And a data formatting step of combining the LZW coded data stream and the lossy compressed coded data stream to generate a coded data stream for the entire image as 'coded image data'. Image compression coding method. 12. The method of claim 11, wherein TH1 is a value between 0.3 and 0.5, TH2 is a value between 0.2 and 0.4, and TH3 is a value of 0.2 or less. Image compression coding method. In the image compression coding system, A pixel to be encoded in an image is set to a 'current pixel P' according to a predetermined scan order, and 'a plurality of predetermined pixels among scan pixels within a predetermined distance from P in the image' (hereinafter, 'compare target pixel'). A search area consisting of &quot; (T) &quot;, and if the pixel value T1 of each T is equal to the pixel value P1 of P, the repeat count value corresponding to the position index D of the corresponding T A same-color pixel search unit for securing [Delta] C data for all Ts with an increment [Delta] C of C) being '1' and otherwise [Delta] C being '0'; Calculate 'C value up to P' (hereinafter referred to as 'CC') by adding △ C to the 'C value up to the previous pixel' (CO) of each T, If MC is greater than 'Maximum value in CO' (hereinafter, abbreviated as 'MO'), the current pixel search result table is generated and provided by setting the CC of each T plus △ C to CC. If MC is less than or equal to MO, it provides 'location index value in search area' (D1) of MO and corresponding T, and initializes all COs to '0', and T with ΔC greater than or equal to 1 after performing the initialization. If is not present, P1 is provided, and ΔC is added to CO of each T, and the current pixel search result table is generated by providing CC. If T with ΔC equal to or greater than 1 exists, the CO of each T is present. A controller configured to generate the current pixel search result table by setting the CC plus the corresponding? C; A pixel search result table storage unit for storing the current pixel search result table; And a basic encoder configured to generate and provide a value obtained by encoding the 'MO and D1' as 'basic encoded data' or to generate and provide the 'encoded value of P1' as basic encoded data. Image Compression Coding System. The method of claim 15, wherein the basic encoder performs basic encoding on 'J current pixels' (hereinafter, referred to as 'J original pixel data columns') to provide a 'basic encoded data stream', wherein J is 8, Any one of 16, 32, 64 Image Compression Coding System. The method of claim 15, The predetermined scan order may be any one of a raster scan order, a left pixel width priority scan order, and a block-by-block raster scan order. Image compression coding method. The method of claim 16, By comparing the data size of the 'J original pixel data columns' (hereinafter referred to as 'A') and the data size of the 'basic encoded data column' (hereinafter referred to as 'B1'), Y1 (= (B1 / A A first selector which provides the basic encoded data string if ()) is less than a first predetermined reference value TH1 (which is a value between 0 and 1) and provides a 'lossy compression encoding command' if Y1 is equal to or greater than TH1. Wow; The apparatus further includes a queue buffer index encoder for generating a queue buffer index encoded data sequence by performing queue buffer index encoding on the basic encoded data sequence. Image Compression Coding System. The method of claim 18, By comparing the data size of the 'cue buffer index coded data stream' (hereinafter referred to as 'B2') and A, Y2 (= (B2 / A)) is a predetermined second reference value TH2 (between 0 and 1). A second selector for providing the queue buffer index coded data string if smaller than TH1, and for providing a 'lossy compression coding command' if Y2 is equal to or larger than TH2; A pixel value difference encoder for generating a pixel value difference encoded data string by performing pixel value difference encoding on the queue buffer index encoded data string; By comparing the data size of the 'pixel value difference coded data stream' (hereinafter referred to as 'B3') and A, Y3 (= (B3 / A)) is a predetermined third reference value TH3 (between 0 and 1). A third selector for providing the pixel value difference encoded data string if smaller than TH2, and for providing a 'lossy compression encoded instruction' if Y3 is equal to or larger than TH3; The apparatus may further include a lossy compression encoding channel configured to perform lossy compression encoding on the J original pixel data sequences in response to the lossy compression encoding command to generate a lossy compression encoded data sequence. Image Compression Coding System. 20. The method of claim 19, wherein the lossy compression coding channel is A DCT unit configured to generate DCT coded data by performing DCT coding on the J original pixel data streams for each block; A quantization unit configured to generate quantized data by performing quantization on the DCT coded data; A Huffman encoder configured to perform Huffman encoding on the quantized data to generate the lossy compressed encoded data stream. Image Compression Coding System. The method of claim 19, And a data formatting unit for generating the encoded image data by combining the lossless compressed coded data string generated for the entire image and the lossy compressed coded data string generated for the entire image. Image Compression Coding System. The method of claim 19, LZW encoding is performed on the 'lossless compressed coded data streams' generated for the entire image to generate LZW encoded data streams for the entire image and combined with the lossy compressed coded data streams generated for the entire image to encode the encoded images. Further comprising a data formatting unit for generating data Image Compression Coding System.