KR100916535B1 - Recorded medium having program for coding and decoding using Bit Precision, and Apparatus and Method there of - Google Patents

Abstract

Disclosed is a recording medium on which a decision bit encoding / decoding apparatus, method, and program according thereto are recorded.

The bit precision encoding apparatus may include: a prediction unit to determine a prediction method of each target pixel of the image data to be encoded, and to calculate a prediction value of the target pixel using the prediction method; An error value and a sine bit are calculated using the prediction value, and the size of the block is sequentially increased from a size of a predetermined block by a predetermined unit, and the first decision bit value corresponding to the error value corresponding to the block is calculated. An encoder which selects and encodes the error value; And a variable block configuration unit configured to calculate variable block configuration information indicating a block configuration in which the compression capacity is minimized using the encoding result of the block. According to the present invention, it is possible to select an appropriate predictor without generating additional bits, and thus the compression efficiency is improved than before.

Decision Bits, Bit Precision, Encoding, Decoding, Complexity

Description

Apparatus and method for determining bit encoding / decoding, and recording medium recording a program according thereto {Recorded medium having program for coding and decoding using Bit Precision, and Apparatus and Method there of}

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus for coding and decoding (hereinafter referred to as 'decoding / decoding'), and more particularly, to encoding / decoding using bit precision in a method of compressing image data. An apparatus, a method and a recording medium having recorded thereon the program are provided.

 As the amount of data transmission per second over the Internet increases, the amount of multimedia data such as video and photos is increasing compared to the past when simple text data was transmitted. Accordingly, there is an increasing demand for image compression technology.

Unlike text data, video data has less redundancy. Compression with a simple encoder shows a compression ratio that is almost the same as the original image capacity. Due to the difference between the image data and the text data, image compression is applied to a prediction technique.

A prediction technique is a technique of predicting a value of a target pixel by using a correlation of a value of one pixel in an image with neighboring adjacent pixel values. When the difference between the predicted value and the actual pixel value is obtained, the value is smaller than that of the original image data and shows a large redundancy feature, thereby increasing the compression efficiency.

Image compression is divided into two types. One is lossless compression, which improves the compression efficiency by removing data that does not show much to the human eye from the original image data, and the other is lossless compression with no loss of the original image data.

Lossless compression has the advantage of maintaining the image quality as it is a compression method in which the original image and the reconstructed image are completely matched, but the compression efficiency is inferior.

Also, with the advent of various mobile media devices in recent years, different requirements from existing video compressors have arisen. Since mobile media devices require limited power and are made of low performance hardware, the execution time for the encoding / decoding process should be short. However, various methods have been published to increase the compression efficiency of image compression, but there is a problem in that the method of reducing the decoding complexity is insufficient.

In order to solve the conventional problems as described above, the present invention proposes an encoding / decoding apparatus, a method, and a recording medium on which the program is recorded, by which an appropriate predictor can be selected without generating additional bits. will be.

In addition, the present invention proposes a decoding apparatus, a method and a recording medium on which the program is recorded, which reduces the complexity and reduces the decoding time in the decision bit decoding method.

Still other objects of the present invention will be readily understood through the following description of the embodiments.

In order to achieve the above object, according to an aspect of the present invention, in the encoding device using a bit precision, the prediction method of each target pixel of the image data to be coded is determined, and the prediction method is used. A predicting unit calculating a predicted value of the target pixel; An error value and a sine bit are calculated using the prediction value, and the size of the block is sequentially increased from a size of a predetermined block by a predetermined unit, and the first decision bit value corresponding to the error value corresponding to the block is calculated. An encoder which selects and encodes the error value; And a variable block configuration unit configured to calculate variable block configuration information representing a block configuration in which a compression capacity is minimized using the encoding result of the block, wherein the encoder is encoded into a block according to the calculated variable block configuration information to form a bitstream. (Bit Stream) is generated, one or more of the variable block configuration information, the sign bit, and the decision bit value are inserted into the bit stream, and the error value and the sign bit are pixel values of the target pixel and the prediction value. Disclosed is a coding apparatus characterized in that the magnitude of the difference and the sign are different.

The prediction unit may determine the prediction method based on whether or not the boundary surface is included in the upper pixel of the target pixel.

The predicted value may be a value calculated by a predetermined method that is equal to or similar to the target pixel value by using a value of a pixel adjacent to the target pixel.

는 하기의 수학식을 이용하여 산출되되, T는 미리 설정된 임계값, A 및 B는 상기 대상 화소의 좌측 및 상측 화소값, E, F, G 및 C는 B에 상응하는 화소를 기준으로 각각 좌상측, 상측, 우상측, 좌측 화소값일 수 있다.The predicted value

Is calculated using the following equation, wherein T is a preset threshold value, A and B are the left and upper pixel values of the target pixel, and E, F, G and C are respectively the upper left on the basis of the pixel corresponding to B. It may be a side, upper side, right upper side, and left pixel value.

[Equation]

인 경우,

If is

이고,

ego,

인 경우,

If is

,

,

및

중 최소값이

,

,

And

Of the minimum

인 경우,

이고,

If is

ego,

인 경우,

이고,

If is

ego,

인 경우,

이고,

If is

ego,

인 경우,

임.

If is

being.

는 에러값,

는 대상 화소의 화소값,

는 대상 화소의 예측값,

는 화소값 및 예측값의 비트수,

는 사인비트,

는 미리 정해진 최대 에러값일 수 있다.The error value and the sine bit are calculated using the following equation,

Is the error value,

Is the pixel value of the target pixel,

Is the predicted value of the target pixel,

Is the number of bits of the pixel value and the predicted value,

Is a sine bit,

May be a predetermined maximum error value.

 [Equation]

이고,

ego,

이면

Back side

The pre-specified block size is 2 ⁿ X2 ⁿ can be (n is a natural number) provided that, n is sequentially increased by one.

The encoder may interpret the sign bit as a sign symbol that is a predetermined unit of bundle of the sign bit.

The encoder may select a second decision bit value corresponding to the first decision bit value, and further encode the first decision bit value using the second decision bit value.

The encoder exchanges the sign of the error value 0 and the sign of the error value 2 ^k -1 when the frequency of the error value 0 among the error values is smaller than the frequency of the error value 2 ^k -1 (k is the first decision bit value). However, a bit indicating whether or not the sign of the error value 0 is exchanged may be inserted into the bitstream.

The encoder compares the frequency of the first decision bit value 1 with the frequency of the remaining first decision bit value among the first decision bit values, and compares the sign of the first decision bit value 1 with the first decision bit value t + 1 ( Where t is a decision bit value that is the largest of the decision bit values having a higher frequency than decision bit value 1), and the sign of the first decision bit values 2 to t is the value of the first decision bit values 1 to t-1. The code may be corrected by a code, but a predetermined 3 bits may be inserted into the bitstream to distinguish the modified code of the first decision bit value 1.

According to another aspect of the present invention, there is provided a decoding apparatus using bit precision, comprising: a block generator for generating a variable block using variable block configuration information extracted from a bit stream to be decoded; A decoder configured to decode the bitstream by using the decision bit value extracted from the bitstream and the variable block configuration information to calculate an error value; And a prediction unit to determine a prediction method of each target pixel of the image data to be decoded, and to calculate a prediction value of the target pixel using the prediction method. And a reconstruction unit configured to calculate a pixel value using the sine bit extracted from the bitstream, the error value, and the prediction value, wherein the error value and the sine bit are equal to the magnitude of the difference between the pixel value of the target pixel and the prediction value. An encoding device is provided, which is a code.

The prediction value may mean a value calculated by a predetermined method that is equal to or similar to the target pixel value by using values of pixels adjacent to the target pixel.

The decoder may decode the encoded first decision bit value using the second decision bit value extracted from the bitstream.

The decoder may decode the error value encoded using the first decision bit value.

The decoding unit extracts a bit indicating a modified position of the first decision bit value 1 among the first decision bit values during encoding in the bitstream, and converts a sign of the first decision bit value 1 into a first decision bit t ( And the sign of the first decision bit value 2 to t to the sign of the corresponding first decision bit value 1 to t-1.

The decoding unit extracts a bit indicating whether the code of the error value 0 during the encoding in the bitstream is exchanged, and if the code is exchanged, the code of the error value 0 and the error value 2 ^k -1 (k May exchange the sign of the first decision bit value).

는 하기의 수학식을 이용하여 산출되되, T는 미리 설정된 임계값, A 및 B는 상기 대상 화소의 좌측 및 상측 화소값, E, F, G 및 C는 B에 상응하는 화소를 기준으로 각각 좌상측, 상측, 우상측, 좌측 화소값일 수 있다.The predicted value

Is calculated using the following equation, wherein T is a preset threshold value, A and B are the left and upper pixel values of the target pixel, and E, F, G and C are respectively the upper left on the basis of the pixel corresponding to B. It may be a side, upper side, right upper side, and left pixel value.

[Equation]

인 경우,

If is

이고,

ego,

인 경우,

If is

,

,

및

중 최소값이

,

,

And

Of the minimum

인 경우,

이고,

If is

ego,

인 경우,

이고,

If is

ego,

인 경우,

이고,

If is

ego,

인 경우,

임.

If is

being.

는 에러값,

는 대상 화소의 화소값,

는 대상 화소의 예측값,

는 화소값 및 예측값의 비트수,

는 사인비트,

는 미리 정해진 최대 에러값일 수 있다.The pixel value is calculated using the following equation,

Is the error value,

Is the pixel value of the target pixel,

Is the predicted value of the target pixel,

Is the number of bits of the pixel value and the predicted value,

Is a sine bit,

May be a predetermined maximum error value.

[Equation]

이면,

If,

According to still another aspect of the present invention, there is provided a coding method using bit precision, comprising: (a) determining a prediction method corresponding to each target pixel of image data to be encoded; (b) calculating a predicted value of the target pixel corresponding to the prediction method; (c) calculating an error value and a sine bit using the predicted value, sequentially increasing the size of the block by a predetermined unit from a size of a predetermined block, and determining a corresponding value to the error value corresponding to the block. Selecting a bit value; (d) encoding the error value using the predetermined decision bit value; (e) calculating variable block configuration information for minimizing a compression capacity by using the encoded result of the block; And (f) encoding a block according to the variable block configuration information to generate a bit stream, wherein the variable block configuration information, the sign bit, and the determination bit value are inserted into the bit stream. And the error value and the sine bit are the magnitude and the sign of the difference between the pixel value of the target pixel and the prediction value.

In step (a), the prediction method may be determined as to whether or not the boundary surface is included in the upper pixel of the target pixel.

The prediction value may mean a value calculated by a predetermined method that is equal to or similar to the target pixel value by using values of pixels adjacent to the target pixel.

는 하기의 수학식을 이용하여 산출되되, T는 미리 설정된 임계값, A 및 B는 상기 대상 화소의 좌측 및 상측 화소값, E, F, G 및 C는 B에 상응하는 화소를 기준으로 각각 좌상측, 상측, 우상측, 좌측 화소값일 수 있다.The predicted value

Is calculated using the following equation, wherein T is a preset threshold value, A and B are the left and upper pixel values of the target pixel, and E, F, G and C are respectively the upper left on the basis of the pixel corresponding to B. It may be a side, upper side, right upper side, and left pixel value.

 [Equation]

인 경우,

If is

이고,

ego,

인 경우,

If is

,

,

및

중 최소값이

,

,

And

Of the minimum

인 경우,

이고,

If is

ego,

인 경우,

이고,

If is

ego,

인 경우,

이고,

If is

ego,

인 경우,

임.

If is

being.

는 에러값,

는 대상 화소의 화소값,

는 대상 화소의 예측값,

는 화소값 및 예측값의 비트수,

는 사인비트,

는 미리 정해진 최대 에러값인 것을 특징으로 하되, 상기 에러값과 사인값은 상기 대상 화소의 화소값과 상기 예측값의 차이의 크기와 부호일 수 있다.The error value and the sine bit are calculated using the following equation,

Is the error value,

Is the pixel value of the target pixel,

Is the predicted value of the target pixel,

Is the number of bits of the pixel value and the predicted value,

Is a sine bit,

Is a predetermined maximum error value, wherein the error value and the sine value may be the magnitude and sign of the difference between the pixel value of the target pixel and the prediction value.

[Equation]

이고,

ego,

이면

Back side

The pre-specified block size is 2 ⁿ X2 ⁿ (n is a natural number) provided that, n this piece can be increased by one in sequence.

In step (c), the sign bit may be interpreted as a sign symbol which is a predetermined unit of bundle of the sign bit.

Step (c) may include (g1) selecting a first decision bit value corresponding to the error value.

Step (c) may further include (g2) selecting a second decision bit value corresponding to the first decision bit value after step (g1).

In the step (d), when the frequency of the error value 0 among the error values is smaller than the frequency of the error value 2 ^k -1 (k is the first decision bit value), the sign of the error value 0 and the sign of the error value k are exchanged. However, a bit indicating whether or not the sign of the error value 0 is exchanged may be inserted into the bitstream.

In step (d), the frequency of the first decision bit value 1 among the first decision bit values is compared with the frequency of the remaining first decision bit values, and the sign of the first decision bit value 1 is converted into the first decision bit value t. Is corrected to a sign of +1 (where t is a decision bit value that is the largest of the decision bit values having a higher frequency than decision bit value 1), and the sign of the first decision bit values 2 to t is the first decision bit value 1 to t Modified by a sign of -1, a predetermined 3 bits may be inserted into the bitstream to distinguish the modified sign of the first decision bit value 1.

According to still another aspect of the present invention, there is provided a decoding method using bit precision, comprising: (a) generating a variable block by extracting variable block configuration information from a bit stream to be decoded; (b) extracting a decision bit value from the bitstream; (c) decoding the bitstream using the decision bit value and the variable block to calculate an error value; (d) determining a prediction method of each target pixel of the image data to be decoded, and calculating a prediction value of the target pixel using the prediction method; (d) restoring image data by using a sine bit extracted from the bitstream, the error value, and the prediction value, wherein the error value and the sine bit are a difference between a pixel value of the target pixel and the prediction value. There is provided a decoding method characterized in that the magnitude and sign of.

The prediction value may mean a value calculated by a predetermined method that is equal to or similar to the target pixel value by using a value of a pixel adjacent to the target pixel.

Step (b) includes (e1) extracting a first decision bit value corresponding to the error value from the bitstream, and (c) step (f1) using the first decision bit value. And decoding the encoded error value.

Step (b) further includes extracting the second bitstream from the bitstream before step (e1), and step (c) uses the second decision bit value before step (f1). And decoding the encoded first decision bit value.

Step (c) extracts a bit indicating whether or not the code of error value 0 is exchanged during encoding in the bitstream, and if the code is indicated to be exchanged, the code of error value 0 and error value 2 among the error values. The sign of ^k -1 (k is the first decision bit value) can be exchanged.

Step (c) extracts a bit indicating a modified position of a first decision bit value 1 of the first decision bit values during encoding in the bitstream, and first determines a sign of the first decision bit value 1. The sign of the bit value t (the modified position) can be exchanged and the sign of the first decision bit value 2 to t can be modified to the sign of the corresponding first decision bit value 1 to t-1.

는 하기의 수학식을 이용하여 산출되되, T는 미리 설정된 임계값, A 및 B는 상기 대상 화소의 좌측 및 상측 화소값, E, F, G 및 C는 B에 상응하는 화소를 기준으로 각각 좌상측, 상측, 우상측, 좌측 화소값일 수 있다.The predicted value

Is calculated using the following equation, wherein T is a preset threshold value, A and B are the left and upper pixel values of the target pixel, and E, F, G and C are respectively the upper left on the basis of the pixel corresponding to B. It may be a side, upper side, right upper side, and left pixel value.

 [Equation]

인 경우,

If is

이고,

ego,

인 경우,

If is

,

,

및

중 최소값이

,

,

And

Of the minimum

인 경우,

이고,

If is

ego,

인 경우,

이고,

If is

ego,

인 경우,

이고,

If is

ego,

인 경우,

임.

If is

being.

는 에러값,

는 대상 화소의 화소값,

는 대상 화소의 예측값,

는 화소값 및 예측값의 비트수,

는 사인비트에 따라 1 또는 -1을 가지는 사인값,

는 미리 정해진 최대 에러값인 것을 특징으로 하되, 상기 에러값과 사인비트는 상기 대상 화소의 화소값과 상기 예측값의 차이의 크기와 부호일 수 있다.The pixel value is calculated using the following equation,

Is the error value,

Is the pixel value of the target pixel,

Is the predicted value of the target pixel,

Is the number of bits of the pixel value and the predicted value,

Is a sine value with 1 or -1 depending on the sine bit,

Is a predetermined maximum error value, wherein the error value and the sine bit may be the magnitude and sign of the difference between the pixel value of the target pixel and the prediction value.

 [Equation]

이면,

If,

According to another aspect of the present invention, a program of instructions that can be executed by an encoding apparatus is tangibly implemented to perform an encoding method using bit precision, and a program that can be read by the encoding apparatus is implemented. A recording medium, comprising: (a) determining a prediction method corresponding to each target pixel of image data to be encoded; (b) calculating a predicted value of the target pixel corresponding to the prediction method; (c) calculating an error value and a sine bit using the prediction value, sequentially increasing the size of the block by a predetermined unit from a size of a predetermined block, and determining bits corresponding to the error value corresponding to the block. Selecting a value; (d) encoding the error value using the predetermined decision bit value; (e) calculating variable block configuration information for minimizing a compression capacity by using the encoded result of the block; And (f) generating a bit stream by encoding the block according to the variable block configuration information, wherein the variable block configuration information, the sign bit, and the decision bit value are inserted into the bit stream. And the error value and the sine bit are magnitudes and signs of the difference between the pixel value of the target pixel and the prediction value.

,

According to another aspect of the present invention, a program of instructions that can be executed by a decoding apparatus is tangibly implemented to perform a decoding method using bit precision, and a program that can be read by the decoding apparatus is implemented. A recording medium comprising: (a) extracting variable block configuration information from a bit stream to be decoded to generate a variable block; (b) extracting a decision bit value from the bitstream; (c) decoding the bitstream using the decision bit value and the variable block to calculate an error value; (d) determining a prediction method of each target pixel of the image data to be decoded, and calculating a prediction value of the target pixel using the prediction method; (d) restoring image data by using a sine bit extracted from the bitstream, the error value, and the prediction value, wherein the error value and the sine bit are a difference between a pixel value of the target pixel and the prediction value. Provided is a recording medium on which a program is recorded, the size of which is and the sign.

The encoding / decoding apparatus, method and recording medium having recorded thereon the program according to the present invention have an advantage that the compression efficiency can be improved compared to the conventional one by selecting an appropriate predictor without generating additional bits.

In addition, in the decision bit decoding method according to the present invention, there is an advantage that the complexity is reduced and the decoding time is reduced.

The above-described embodiments of the present invention are disclosed for the purpose of illustration, and those skilled in the art may make various modifications, changes, and additions within the spirit and scope of the present invention. Should be considered to be within the scope of the following claims.

As the invention allows for various changes and numerous embodiments, particular embodiments will be illustrated in the drawings and described in detail in the written description. However, this is not intended to limit the present invention to specific embodiments, it should be understood to include all transformations, equivalents, and substitutes included in the spirit and scope of the present invention. In the following description of the present invention, if it is determined that the detailed description of the related known technology may obscure the gist of the present invention, the detailed description thereof will be omitted.

Terms such as first and second may be used to describe various components, but the components should not be limited by the terms. The terms are used only for the purpose of distinguishing one component from another.

As the invention allows for various changes and numerous embodiments, particular embodiments will be illustrated in the drawings and described in detail in the written description. However, this is not intended to limit the present invention to specific embodiments, it should be understood to include all modifications, equivalents, and substitutes included in the spirit and scope of the present invention. In the following description of the present invention, if it is determined that the detailed description of the related known technology may obscure the gist of the present invention, the detailed description thereof will be omitted.

Terms such as first and second may be used to describe various components, but the components should not be limited by the terms. The terms are used only for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, the first component may be referred to as the second component, and similarly, the second component may also be referred to as the first component. The term and / or includes a combination of a plurality of related items or any item of a plurality of related items.

When a component is referred to as being "connected" or "connected" to another component, it may be directly connected to or connected to that other component, but it may be understood that other components may be present in between. Should be. On the other hand, when a component is said to be "directly connected" or "directly connected" to another component, it should be understood that there is no other component in between.

The terminology used herein is for the purpose of describing particular example embodiments only and is not intended to be limiting of the present invention. Singular expressions include plural expressions unless the context clearly indicates otherwise. In this application, the terms "comprise" or "have" are intended to indicate that there is a feature, number, step, operation, component, part, or combination thereof described in the specification, and one or more other features. It is to be understood that the present invention does not exclude the possibility of the presence or the addition of numbers, steps, operations, components, components, or a combination thereof.

Unless defined otherwise, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art. Terms such as those defined in the commonly used dictionaries should be construed as having meanings consistent with the meanings in the context of the related art, and are not construed in ideal or excessively formal meanings unless expressly defined in this application. Do not.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the following description of the present invention, the same reference numerals will be used for the same means regardless of the reference numerals in order to facilitate the overall understanding.

First, to help understand the encoding scheme (hereinafter, referred to as 'decision bit encoding') using the decision bit value according to the present invention, the concept will be described.

The decision bit encoding method according to the present invention is a method of determining a fixed bit value with respect to input data symbol values using, for example, a cost function, and binarizing the symbol value using bits as large as the value. For example, assume that the input symbol values are {2, 1, 0, 4, 5, 3}, and the decision bit value is 2 bits. A value that can be represented by two bits can represent a value from 0 to 3. However, if the symbol value is larger than 3, two bits cannot be represented, and thus, additionally two more bits are allocated to represent a value larger than three. More detailed description is as follows. The symbol value 2 is represented by '10' when the value is binarized as it is, the symbol value '1' is represented by '01' when it is binarized, and the symbol value '0' is represented by '00' when it is binarized. However, since the symbol value '4' is larger than 3, which is the maximum value represented by 2 bits, the symbol value '4' cannot be represented by only 2 bits. Therefore, 3, the maximum value that can be represented by the first two bit values, is represented by '11', and insufficient value 1 is represented by '01', thereby connecting two results. That is, the symbol value '4' may be expressed as '1101'. Similarly to the symbol value 4, the symbol value 5 is expressed as '1110' by combining the maximum value 3 and the surplus value 2 using 2 bits. In the case of symbol value 3, 2 bits can be expressed. However, in this case, it is indistinguishable from the case of expressing a value larger than 3, and additional 2 bits are added so that the surplus value is 0. Will be represented.

If the above-described decision bit coding scheme is applied to a case where the symbol value of data is 0 to 9, for example, it may be represented as shown in Table 1 below.

Decision bit value 1 bit Decision bit value 2 bit Decision bit value 3 bit Decision bit value 4 bit Decision bit value 5 bits 0 0 00 000 0000 00000 One 01 01 001 0001 00001 2 001 10 010 0010 00010 3 0001 1100 011 0011 00011 4 00001 1101 100 0100 00100 5 000001 1110 101 0101 00101 6 0000001 111100 110 0110 00110 7 00000001 111101 111000 0111 00111 8 000000001 111110 111001 1000 01000 9 0000000001 11111100 111010 1001 01001

By encoding the input data using the decision bits as described above, not only the encoding of the input data can be performed, but also the compression performance similar to that of the conventional Hoffman coding method can be obtained with little calculation.

1 is a block diagram of an encoding apparatus according to an embodiment of the present invention, and FIG. 2 is a flowchart of an encoding method according to an embodiment of the present invention.

Referring to FIG. 1, the decision bit encoding apparatus 100 according to an exemplary embodiment of the present invention may include a data input unit 110, a prediction unit 120, an encoding unit 130, a variable block construction unit 140, and a bitstream. (bit stream) may include a transmission unit 150.

Hereinafter, in order to facilitate understanding and explanation of the invention, the functions and operations of the components of FIG. 1 will be described with reference to a flowchart of an encoding method according to an embodiment of the present invention of FIG. 2.

Referring to FIG. 2, in step S210, the data input unit 110 receives image data from an external device.

In operation S220, the prediction unit 120 calculates a prediction value of the target pixel to be predicted from the input image data (hereinafter, referred to as an “input image”), and then calculates a difference between the prediction value and the pixel value of the target pixel. The error value and the sign bit of the magnitude and the sign are calculated. This will be described later in detail with reference to FIGS. 7 to 9.

Then, the encoding unit 130 2 ⁿ X2 ⁿ in step S230 (n is a natural number of 1 or more) coded for by the error value described above by using the bit-precision value selected in units of blocks of a size and compression of each block in the encoding capacity Save it. In addition, the encoder 130 sequentially increases the value of n until the larger block cannot be set, repeats the above-described encoding process in units of blocks, and stores the compression capacity.

In operation S240, the variable block configuring unit 140 configures the variable block having the smallest compression capacity of the entire input image by using the compression capacity stored in operation S230. This will be described later in detail with reference to FIG. 3.

In addition, although not shown in FIG. 2, according to another embodiment of the present invention, in step S230, the encoder 130 determines a decision bit value selected in units of blocks of size 2 ⁿ X 2 ⁿ (n is a natural number of 1 or more). Encode by using and store the compressed capacity, but specify n as 1 in the first encoding step. In step S240, the variable block configuration unit 140 generates variable block configuration information indicating the current variable block configuration. Subsequently, the encoder 130 encodes each block by specifying n as 2. The variable block constitution unit 140 is configured to sum the compressed capacities of blocks including the same pixel as each block when n is set to 2 among blocks currently configured in the variable block, when n is designated as 2. Reconstruct the variable block using smaller blocks compared to the compressed capacity of each block. Re-encoding the bit-precision unit 100 repeats S230 and S240 until the 2 ⁿ X2 ⁿ block size of not more than the size of the image.

Subsequently, in step S250, the variable block configuration unit 140 generates variable block configuration information indicating the variable block configured in step S240. The variable block configuration information is expressed in a bit string of a predetermined format and inserted into the bitstream.

The decision bit coding method according to the embodiment of the present invention has been described so far. The steps S230 to S270 described above will be described in more detail with reference to FIG. 3 and FIG. 4, taking an example in which the input image is an 8 × 8 image.

FIG. 3 is a diagram illustrating a unit block for explaining a variable block configuration, and FIG. 4 is a diagram illustrating a quadtree generated when the input image of FIG. 3 is encoded according to an embodiment of the present invention.

3, the size of the smallest block (e.g. 310) of the block shown in Figure 3 is a 2 ¹ X2 ^1. Variable block generating unit 140 divides the image into a 2 ¹ X2 ¹ block and create a variable block configuration information indicating a configuration of the divided blocks. In this case, the variable block configuration information may be represented by a quadtree. In addition, it will be apparent to those skilled in the art that the variable block configuration information may be expressed in a format other than the quadtree. The initial quadtree is also a full quadtree with 21 nodes. The encoder 130 performs encoding in units of blocks and temporarily stores the compression capacity of each block.

Subsequently, the variable block configuration unit 140 divides the image into 2 ² X ^{2 2} blocks (for example, 350), and the encoder 130 performs encoding for each block. The sum of the compression capacity of a variable block configuration section 140 is a ² 2 X2 2 block 350 compression capacity and 2 ^1, X2 of the stored block is compressed to ¹ block (i.e., block 310, 320, 330 and 340) of the encoded Compare.

In describing the present invention, it is assumed that the sum of the compression capacities of blocks 310, 320, 330, and 340 has a smaller capacity than the compression capacities of 2 ² X ^{2 2} blocks 350 in order to facilitate an overall understanding. The variable block construction unit 140 temporarily stores the compression capacities of the blocks 310, 320, 330, and 340, and modifies the quadtree to represent this block configuration. For example, the variable block configuration unit 140 uses the blocks 310 to 340 instead of the block 350 in the current step by keeping the nodes 410 to 440 of FIG. 4 corresponding to the blocks 310 to 340 of FIG. 3 in the quadtree. Express that.

Subsequently, the variable block configuration unit 140 repeats the same process for each of blocks 360, 370, and 380.

Here, it is assumed that the compression capacities of the 2 ² X ^{2 2} blocks 360, 370, and 380 are smaller than the sum of the compression capacities of the 21 × 21 units included in each. In this case, the variable block configuration unit 140 temporarily stores the compression capacities of 2 ² X ^{2 2} blocks 360, 370, and 380. In addition, the variable block configuration unit 140 expresses the use of blocks 360 to 380 by removing the lower nodes of the nodes 460 to 480 of FIG. 4.

The variable block configuration section 140, executes the encoding by 2 ³ X2 ³ blocks, such as image size, and the temporarily stored in block 310, 320, 330, 340, 360, 370 and combined with the block 390 of 380 compressing capacity of the coding Compare the compression capacity. Here, assuming that the compression capacity of the block 390 is larger, the variable block configuration unit 140 selects blocks 310, 320, 330, 340, 360, 370, and 380 as blocks for encoding. The encoder 130 performs final encoding using blocks 310, 320, 330, 340, 360, 370, and 380. In this case, since the variable block configuration unit 140 has a larger compression capacity of the block 390, as shown in FIG. 4, the final quadtree configured without removing the lower node of the node 490 is a bitstream in a bit representation of a predetermined format. Add to

Nodes 410 through 480 shown in FIG. 4 correspond to blocks 310 through 390 shown in FIG. That is, the variable block configuration unit 140 sequentially encodes the block while increasing the size of the block, selects a convex size unit having a small compression capacity, and modifies the quadtree accordingly.

Here, when the size of the image data will not fall into the 2 ⁿ X2 ⁿ block may specify 2 ⁿ X2 ⁿ different size of the block as a separate method, it is obvious to those skilled in the art.

In addition, in the above description, the quadtree (see FIG. 4) is modified according to a process in which the variable block configuration unit 140 selects a size unit of a block having a small compression capacity from an input image. However, according to another embodiment of the present invention, after the variable block configuration unit 140 selects a size unit of a block having a small compression capacity from the input image, the quadtree may be finally modified.

5 is a diagram illustrating dividing a Lena image into blocks having a size of 64 × 64.

Referring to FIG. 5, the Lena image is image data having 512 × 512 resolution. In this case, when the size information of all the designated blocks is 64 × 64, as illustrated in FIG. 5, the image data may be divided into eight rows and eight columns. That is, the total number of blocks can be 64.

In the following embodiment of the present invention, a case where a Lena image is divided as illustrated in FIG. 5 will be described as an example. Although all blocks are designated as 64 X 64 in FIG. 5, this is for convenience of understanding and explanation of the present invention. In the present invention, the variable block configuration unit 140 is described with reference to FIGS. 2 to 4. , The size of each block is specified flexibly.

Hereinafter, the DPCM (Differential Pulse Code Modulation) scheme will be briefly described with reference to FIG. 6. In addition, we look at the limitations of the DPCM method. Subsequently, the operation of calculating the predicted value by predicting the pixel value of the image data by the predictor 120 according to an exemplary embodiment of the present invention (S220) will be described in detail with reference to FIGS. 7 to 9.

6 is a diagram for explaining a DPCM.

The DPCM encodes a prediction error that is a difference between the predicted pixel value and the pixel value of the target pixel. The DPCM is a predictive encoding method in which the pixel of the target pixel is reconstructed by the sum of the predicted value and the error value. DPCM is a simple technique that is commonly used and easy to implement. The predicted error value maintains a reduced deviation because the prediction uses a local correlation of neighboring neighboring pixel values. Therefore, when using the prediction technique, it is possible to encode using fewer bits when compared to directly encoding the original pixel value. In this way, when the error image (the image in which each pixel is expressed as an error value) is compressed, the compression efficiency is higher than when the original image is compressed. Since DPCM alone does not obtain data compression, only the statistical properties change, it is combined with an encoding technique that usually serves as data compression. The fundamental characteristic of DPCM is that the predictable components are automatically reproduced by the decoding apparatus, and only the unpredictable components are encoded and sent. Therefore, when the current value can be predicted more closely from the past values, the compression rate is also higher when the coding technique is used.

Referring to FIG. 6, the predicted value may be expressed by a relational expression of the previous peripheral pixel values A, B, C, D, and E. The DPCM method can easily predict the pixel value X. However, due to the characteristics of the DPCM, a large error occurs in an edge portion where the correlation between pixels is small. In addition, even if the same image is partially different from each other, there is a limit in properly predicting all pixel values in the image with one predictor.

In the prior art, in order to further improve the prediction technique described above, for example, an additional bit is used to convert to a more suitable bit representation. However, this method has a problem in that compression efficiency is lowered due to capacity increase due to additional bits.

In the embodiment of the present invention, the compression efficiency is increased by making prediction without using additional bits. More specifically, it is used that the prediction direction of the target pixel having an interface in the image data is mostly similar to the prediction direction of the upper pixel. Therefore, in the embodiment of the present invention, the prediction direction of the upper pixel of the target pixel to be predicted is applied to the target pixel. This point will be described below with reference to FIGS. 7 to 9.

7 and 8 are diagrams for describing a method of predicting each pixel value of image data according to an embodiment of the present invention, and FIG. 9 is a prediction algorithm considering an interface of image data according to an embodiment of the present invention. to be.

7, the position of each pixel is briefly described. FIG. 7 is a diagram illustrating that each pixel is disposed on image data. In more detail, when the target pixel to be predicted is X (i, j), it may be assumed to be X 710 of FIG. 7. Where i is the position of the row of pixels and j is the position of the column.

In this case, the left pixel X (i, j-1) of the target pixel X (i, j) to be predicted may be represented by A 711 of FIG. 7. In the same manner, the upper left pixel X (i-1, j-1), the upper pixel X (i-1, j) and the upper right X (i-1, j) of the target pixel X (i, j) to be predicted +1) may be represented by C 713, B 712, and D 714 of FIG. 7, respectively. Hereinafter, the structure of the pixel illustrated in FIG. 7 will be described as an example to facilitate understanding and explanation of the invention.

In the image data, the portion having the boundary surface may have the same prediction direction as that of the immediately upper pixel. The present invention takes advantage of this feature. That is, in order to predict the pixel value of the target pixel, the optimum prediction direction of the upper pixel is used.

The prediction method according to an embodiment of the present invention predicts using Equation 1 below when no boundary exists in the pixel X 710 to be predicted. This prediction method does not use additional bits, and the prediction process is simple and causes little complexity.

 [Equation 1]

In this case, in order to determine whether or not the boundary exists in the pixel X 710 to be predicted, a determination as to whether or not the boundary exists in the upper pixel B 712 is used. More specifically, whether the upper pixel B 712 has a boundary as to whether or not the difference between the pixel values of the left pixel C 713 and the right pixel D 714 of the upper pixel B 712 is equal to or less than a predetermined threshold value. Can be determined. For example, looking at an enlarged view of a portion of the Lena image shown in FIG. 8, there is a boundary of the image. The difference between the pixel C 713 and the pixel D 714 at this interface has a larger value than the difference between the pixels without the interface. Using this point, the prediction unit 120 may determine the boundary surface when the difference between the C 713 and the D 714 of the upper pixels is greater than or equal to a predetermined value in order to predict the target pixel X 710.

That is, when the difference between the pixel value of the left pixel C 713 and the right pixel D 714 of the upper pixel B 712 of FIG. 7 is equal to or less than a predetermined threshold value, no boundary surface exists in the upper pixel B 712. You can judge that you do not. In this case, as described above, the pixel value of the pixel X 710 may be predicted as Equation 1 above.

However, when the difference between the pixel values of the left pixel C 713 and the right pixel D 714 of the upper pixel B 712 is equal to or more than a predetermined threshold value, it can be determined that an interface exists in the upper pixel 712. have. In this case, the probability that an interface exists in the pixel X 710 increases. In this case, the prediction method according to the embodiment of the present invention finds the boundary on the image data and determines the prediction method.

8, the direction of the boundary surface of the pixel X 710 and the direction of the boundary surface of the pixel B 712 show a similar tendency. Using this point, the prediction direction determined by determining whether or not an upper boundary of the upper pixel B 712 exists may be applied to the pixel X 710. More specifically, by using adjacent pixels (eg, C 713, E 715, F 716, and G 717) of the upper pixel B 712, the pixel B 712 is formed. The optimal prediction direction may be determined and used as the prediction direction of the target pixel X 710 to be predicted. If the prediction direction of the upper pixel is used as it is, the complexity may be caused in this case, but there is an advantage in that compression efficiency is increased because additional bits indicating the prediction direction are not added to the bitstream.

It is apparent to those skilled in the art that the algorithm illustrated in FIG. 9 can be used as an algorithm for such a prediction method according to an embodiment of the present invention. In addition, it will be apparent to those skilled in the art that the algorithm may be variously changed or modified.

In this specification, the prediction direction of the target pixel is determined by using the prediction direction of the upper pixel of the target pixel to be predicted. However, it will be apparent to those skilled in the art that this method is only an embodiment of the present invention, and that the prediction direction of several pixels close to the target pixel can be used.

Up to now, the prediction method according to the embodiment of the present invention has been described. Hereinafter, a method of resetting the prediction value calculated by the prediction unit 120 through the above-described prediction method according to an embodiment of the present invention will be described with reference to FIG. 10.

FIG. 10 is a diagram illustrating resetting a predicted value according to an embodiment of the present invention, and FIG. 11 is a diagram illustrating a distribution of difference values between a target pixel value and a predicted value of image data according to an embodiment of the present invention. .

According to an embodiment of the present invention, the error value E may be calculated using Equation 2 below using the predicted value calculated by the predictor 120.

[Equation 2]

Here, E is an error value, X is a pixel value of the target pixel (that is, a pixel value of X 710 in FIG. 7), and P (X) is a pixel value predicted in the target pixel (ie, a prediction value).

의 부호는 사인비트로서 별도로 처리 될 수 있다. 예를 들어, 부호가 양수인 경우 사인비트가 1로 저장 되고, 부호가 음수인 경우 사인비트가 0으로 저장될 수 있다. 또는 그 반대로 저장될 수 있음은 당업자에게 자명하다. 이하, 부호가 양수인 경우 사인비트가 1로 저장 되고, 부호가 음수인 경우 사인비트가 0으로 저장되는 경우를 가정하여 설명한다.At this time,

The sign of can be treated separately as a sign bit. For example, if the sign is positive, the sign bit may be stored as 1, and if the sign is negative, the sign bit may be stored as 0. It will be apparent to those skilled in the art that the data can be stored or vice versa. Hereinafter, it is assumed that the sign bit is stored as 1 when the sign is positive and the sign bit is stored as 0 when the sign is negative.

In general, when the image data is a gray scale image, the luminance value is represented by 8 bits (ie, 28). In the case of high quality, the luminance value may be represented by 10 bits (ie, 210). In this specification, for convenience of understanding and description of the present invention, it is assumed that the luminance value is 8 bits. However, it will be apparent to those skilled in the art that the present invention is not limited thereto.

은 -255 내지 255의 값이다. 따라서, 상술한 수학식 2를 이용하여 예측부(120)가 산출한 예측값은 0 내지 255사이의 값이다. 즉, 에러값의 최대치가 255가 된다.If the luminance value is 8 bits,

Is a value from -255 to 255. Therefore, the predicted value calculated by the predictor 120 using Equation 2 described above is a value between 0 and 255. That is, the maximum value of the error value is 255.

At this time, in the embodiment of the present invention, as illustrated in FIG. 10, the maximum value of the error value may be reset to 7 bits (that is, up to 127 can be represented).

An embodiment of the present invention is described in more detail with reference to FIG. 10, where the calculation on the vertical line 1010 is calculated on the circle 1020. That is, suppose there is 1 again after 256.

가 30이고, 예측부(120)가 산출한 예측값

이 200인 경우를 가정하자. 이 경우 수학식 2에 의하면 에러값

는 170이 되고,

의 부호는 음수(-)이다. For example, the target pixel value

Is 30, and the prediction value calculated by the prediction unit 120

Assume the case is 200. In this case, according to equation (2)

Becomes 170,

The sign of is negative.

는 86이 되고,

의 부호는 양수가 될 수 있다. 이 경우, 향후 복호화 과정에서 화소값

는 286이 되지만, 최대값인 256보다 큰 값이므로, 256을 뺀 30으로 복호화하게 된다. 상술한 방식으로 에러값을 표현하면 7비트, 즉 127이내의 값으로 표현이 가능하다. However, when viewed on the circle 1020 according to an embodiment of the present invention, an error value

Becomes 86,

The sign of may be positive. In this case, the pixel value in the future decoding process

Becomes 286, but is larger than the maximum value of 256, so it is decoded by 30 minus 256. When the error value is expressed in the above-described manner, it is possible to express the value within 7 bits, that is, 127.

The reason for reducing the error value to 127 or less as described above is to increase the compression efficiency by reducing the number of bits to be encoded because the error value itself is encoded in the encoding using the decision bit value described below.

As shown in FIG. 11, most of the difference between the target pixel value and the predicted value is shown as a value between −20 and +20. Since the higher the frequency of the small error value, the greater the compression efficiency during encoding. Therefore, when the error value is reset by reducing the maximum value according to an embodiment of the present invention, it is apparent to those skilled in the art that the compression rate is high.

Up to now, a method of calculating an error value using luminance values according to an embodiment of the present invention in a gray scale image has been described with reference to FIG. 10. However, it will be apparent to those skilled in the art that the present invention may be applied together with a prediction technique using a relationship between colors in a yuv image or an RGB representation in a color image.

Hereinafter, a method of encoding the error value and / or the sine bit calculated by the above-described method will be described.

12 is a flowchart illustrating a method of encoding by an encoder using decision bit values according to an embodiment of the present invention.

In the present specification, a 512 X 512 Lena image illustrated in FIG. 5 will be described as an example. 5 and 12, in step S1210, the encoder 130 selects a first decision bit value of each block.

Here, in FIG. 12, all the blocks are consistently 64X64 size in order to understand the present invention. However, the variable block configuration unit 140 may select the size of each block differently with reference to FIGS. 2 to 4.

According to an embodiment of the present invention, three bits (ie, 23) may be allocated to the decision bit value. That is, the decision bit value may be 0 to 7.

Here, when the size information of all the designated blocks is 64X64, a total of 4096 pixels are included in one block. That is, the number of error values calculated by the predictor 120 in one block is also 4096. Therefore, according to an embodiment of the present invention, the encoder 130 performs decision bit coding on 0 to 7 for 4096 prediction values in order to select decision bit values corresponding to a predetermined block, and then the compression efficiency is the best. A high number is chosen as the decision bit value for a given block.

For example, an error value predicted for a pixel included in a predetermined block is {5, 3, 8, 2... }. In the following description, a method of selecting a decision bit value using four error values as an example for convenience of understanding and explanation of the invention will be described. The error value is {5, 3, 8, 2... } With respect to the encoder 130, the decision bit encoding may be performed using 0 to 7 as shown in Table 2 below.

Error value 5 3 8 2 Total bits One 000001 0001 000000001 001 22 2 1110 1100 111110 10 18 3 101 011 111001 010 15 4 0101 0011 1000 0010 16 5 00101 00011 01000 00010 20 6 000101 000011 001000 000010 24 7 0000101 0000011 0001000 0000010 28

Referring to Table 2, the use of 3 as the decision bit value has the highest compression efficiency. In this case, the encoder 130 may select 3 as the decision bit value for the predetermined block.

In the above-described method, the encoder 130 may select a decision bit value corresponding to the predetermined block in step S1210.

Here, in order to select a decision bit value corresponding to a predetermined block, after encoding the pixels of a predetermined specific position among the pixels included in the predetermined block, the decision bit value may be selected by examining the compression efficiency. Self-explanatory

Subsequently, in operation S1220, the encoder 130 determines bit-encodes error values included in each block by using the selected first decision bit value.

In addition, a first decision bit value corresponding to each block (for example, decision bit value 3 of block 510 of FIG. 5) may be inserted into the bitstream generated in operation S1230.

At this time, it is apparent to those skilled in the art that the size information of the block and the first decision bit value may be variously inserted into the hierarchical structure of the bitstream according to various embodiments of the present disclosure.

Until now, the encoding unit 130 selects and encodes the decision bit value corresponding to the block designated in the image data (for example, the 512X512 Lena image of FIG. 5) according to the embodiment of the present invention.

According to another embodiment of the present invention, the compression efficiency may be increased by additionally performing decision bit coding on the decision bit value corresponding to the block. This will be described later in detail with reference to FIG. 13.

13 is a flowchart illustrating a method of encoding by an encoder using decision bit values according to another embodiment of the present invention.

In the following, descriptions overlapping with those described with reference to FIG. 12 will be omitted to clarify the gist of another embodiment of the present invention.

In operation S1330, the encoder 130 selects a second decision bit value with respect to the first decision bit value corresponding to each block of the image data.

More specifically, in the case of the Lena image described with reference to FIGS. 5 and 12, a total of 64 blocks are divided according to the designated block size information. Therefore, according to an embodiment of the present invention, a total of 64 first decision bit values are selected corresponding to each block. Here, the encoder 130 selects second decision bit values for the 64 first decision bit values. Since the process of selecting the second decision bit value is the same as that of step S1210 of FIG. 12, duplicate description thereof will be omitted.

At this time, it will be apparent to those skilled in the art that the encoder 130 may select the second decision bit value based on a predetermined bundle unit of the first decision bit value.

Subsequently, in step S1340, the encoder 130 encodes the first decision bit value using the selected second decision bit value.

In this case, when the second decision bit value is selected for the first decision bit value in the predetermined bundle unit, the encoder 130 uses the second decision bit value corresponding to the first decision bit value in the predetermined bundle unit. It will be apparent to those skilled in the art that the first decision bit value of the predetermined packed unit can be determined by bit coding.

Subsequently, in operation S1350, the encoder 130 may insert the second decision bit value into the generated bitstream.

So far, the method of encoding the error value calculated by the encoder 130 by the encoder 120 has been described with reference to FIGS. 12 to 13. Hereinafter, a method of encoding the sine bits calculated by the predictor 120 by the encoder 130 will be described.

의 부호이다. 바람직하게는 본 발명의 실시예에 따라 1비트이다. 따라서, 사인비트는 화소당 1비트로서, 결정 비트 부호화하여도 압축효율이 높지 않다. 본 발명의 실시예에 따르면, 사인 비트를 미리 지정된 묶음단위로 사인심볼로 해석하고, 이를 생성되는 비트스트림에 삽입할 수 있다.As described above, the sine bit is expressed by Equation 2

Is the sign of. Preferably it is 1 bit in accordance with an embodiment of the invention. Therefore, the sine bits are 1 bit per pixel, and the compression efficiency is not high even when the decision bit coding is performed. According to an embodiment of the present invention, the sinusoidal bits may be interpreted as sinusoidal symbols in a predetermined bundle unit and inserted into the generated bitstream.

More specifically, for example, it is assumed that there are 8 predetermined bundle units of sine bits, the image data is a 512 × 512 Lena image of FIG. 5, and the size information of all the designated blocks is 64 × 64.

For example, in a given block (eg, 510 of FIG. 5), if the sign bits of the first and eight pixels are {0, 0, 0, 0, 1, 1, 1, 1}, the sign symbol is Can be interpreted as 7. In this case, since the size information of the block is 64X64, the number of sign symbols may be 512. In this case, the encoder 130 may insert 512 symbols into the generated bitstream.

Thus far, the method for encoding the sine bits calculated by the predictor 120 according to the embodiment of the present invention has been described.

Hereinafter, a method of performing additional code processing by the encoder 130 according to an embodiment of the present invention will be described with reference to FIGS. 14 to 16.

14 is a flowchart illustrating a process of improving compression efficiency by additional code processing to a decision bit encoding method according to an embodiment of the present invention. FIG. 15 is an error value distribution diagram of a Lena image according to an embodiment of the present invention. Table 3 is a table comparing the code before the code exchange and the code after the code exchange of the error value of the Lena image according to an embodiment of the present invention.

Referring to FIG. 15, a distribution of 0 of error values of a Lena image according to an embodiment of the present invention has a frequency smaller than that of error values 1, 2, and 3. Referring back to FIG. 14, in operation S1410, the encoder 130 measures the frequency of the error value calculated by the predictor 120.

Subsequently, in step S1420, the encoder 130 proceeds to step S1430 when the frequency of the error value 0 is smaller than the frequency of the error value b (the arbitrary decision bit value). Otherwise, the encoder 130 proceeds to step S1440.

In step S1430, the encoder 130 exchanges a sign of the error value 0 and a sign of the error value b. In step S1440, one bit may be added for later checking as to whether or not signs are exchanged in the generated bitstream. At this time, it is apparent to those skilled in the art that the bit assignment according to the exchange status may be different depending on the case.

More specifically, assuming that a decision bit value for encoding an error value is 2 through a prediction value, 2 bits should be used for encoding error values 0, 1, and 2, and 3, 4, and 5 are encoded. 4 bits should be used. For example, referring to FIG. 15, the error value is {1, 2, 3, 0, 4, 5... } The frequency is shown in the order shown. Here, the frequency of 0 is interchanged with the symbol '00' used for encoding 0 and the symbol '1100' for expressing 3 using a point smaller than the frequency of 3. The result is a reduction in compression capacity that is twice the frequency difference between zero and three. As a result of the sign exchange, the sign of the error values 0 and 3 is changed as shown in Table 3.

Error value Code before code exchange of error value Code after code exchange of error value 0 00 1100 One 01 01 2 10 10 3 1100 00 4 1101 1101 5 1110 1110 ... ... ...

More generally, when the frequency of the error value 0 is smaller than the frequency of the error value 2 ^k -1 (k is the first decision bit value: for example, 3), the sign of the error value 0 and the error value 2 ^{k are described.} Exchange the sign of -1. Up to now, the replacement of the decision bit value according to the frequency of the error value has been described with reference to FIGS. 14 to 15.

Hereinafter, the exchange of the second decision bit value according to the distribution of the first decision bit value will be described with reference to FIG. 16.

16 is a distribution diagram of decision bit values selected for each block in a Lena image according to an embodiment of the present invention, and Table 4 is a code before code exchange of a first decision bit value of a Lena image according to an embodiment of the present invention. A table comparing codes after code exchange.

Referring to FIG. 16, the decision bit values of the Lena image according to the present invention are concentrated in 2 and 3. As the decision bit value selected for the error value is exchanged as described above, if the second decision bit value selected for the first decision bit value is exchanged, the compression efficiency can be further increased.

The first decision bit value of FIG. 16 gradually decreases in frequency from 2 to 8. FIG. The frequency of the first decision bit value 1 is smaller than that of the first decision bit value 2. Although the distribution of the first decision bit value shows some differences for each image, the first decision bit value is generally similar to that of FIG. 16.

The encoder 130 sequentially compares the frequencies of the first decision bit values 1 and the frequencies of the remaining first decision bit values 2 to 8. For example, when the former decision bit value 1 is compared with the first decision bit value 2 and the former value is smaller, the first decision bit value 1 and the first decision bit value 3 are compared. In the case of the Lena image, when the first decision bit value 1 has a larger value when compared with the first decision bit value 6, the sign of the first decision bit value 1 is corrected to the sign of the first decision bit value 5. . As shown in Table 4, the sign of the first decision bit value 2 to the first decision bit value 5 is modified to the sign of the corresponding first decision bit value 1 to the first decision bit value 4.

First decision bit value Sign of decision bit value before sign exchange Sign of decision bit value after sign exchange One 00 1101 2 01 00 3 10 01 4 1100 10 5 1101 1100 6 1110 1110 7 111100 111100 8 111101 111101

More generally, the encoder 130 may determine the frequency of 1 of the first decision bit values 1 to 8 and the frequency of the remaining first decision bit values 2 to 8 of the first decision bit value 1. After comparing the first decision bit value t (e.g. 6) with a greater frequency of, modify the sign of the first decision bit value t-1 (e.g. 5), The sign of the first decision bit value 2 to t-1 is corrected to the sign of the corresponding first decision bit value 1 to t-2 (for example 4).

So far, the method of encoding the error value calculated by the predicting unit 120 by the encoding unit 130 has been described with reference to FIGS. 12 to 13. In addition, the method in which the encoder 130 encodes the sine bit calculated by the predictor 120 has been described. In addition, with reference to FIGS. 14 to 16, a method in which the encoding unit 130 improves compression efficiency through code exchange between an error value and a decision bit value has been described.

Here, according to the exemplary embodiment of the present invention, it will be apparent to those skilled in the art that the encoder 130 may select and encode the decision bit value in various ways from the error value and / or the sine bit calculated by the predictor 120. .

Hereinafter, an apparatus and method for determining bit decoding according to an embodiment of the present invention will be described.

17 is a configuration diagram of a decision bit decoding apparatus according to an embodiment of the present invention, and FIG. 18 is a flowchart of a decision bit decoding method according to an embodiment of the present invention.

Referring to FIG. 17, a decision bit decoding apparatus according to an embodiment of the present invention includes a bitstream input unit 1710, a variable block generator 1720, a decoder 1730, a predictor 1740, and a reconstructor 1750. It may include.

Hereinafter, in order to facilitate understanding and explanation of the present invention, the functions and operations of the respective transport elements of FIG. 17 will be described with reference to a flowchart of a decoding method according to the embodiment of the present invention of FIG.

Referring to FIG. 18, in step S1810, the bitstream input unit 1710 receives an encoded bitstream.

In operation S1820, the variable block generator 1720 extracts quadtree information from the received bitstream and generates a variable block used for encoding.

Subsequently, in step S1830, the decoder 1730 decodes the error value by using the decision bit value extracted from the bitstream.

Subsequently, in step S1840, the prediction unit 1740 calculates a prediction value using the peripheral pixel values of the target pixel, and the reconstruction unit 1750 calculates the target pixel value using the prediction value, the decoded sine bits, and the error value. Restore the original image.

Hereinafter, the processes of steps S1830 and S1830 will be described in detail with reference to FIG. 19.

19 is a flowchart illustrating a method of decoding a bitstream by a decoder, a predictor, and a reconstructor according to an embodiment of the present invention.

In describing the decision bit decoding method according to an exemplary embodiment of the present invention, a description overlapping with the reverse order of the decision bit encoding method described with reference to FIGS. 12 to 16 will be omitted for clarity.

In addition, since the decoding method using the determination bit value is the reverse order of the above-described determination bit encoding method, a detailed description thereof will be omitted.

First, a method of calculating error values from the bitstream by the decoder 1730 according to an embodiment of the present invention will be described.

In operation S1910, the decoder 1730 extracts a second determination bit value from the bitstream to be decoded, and in operation S1911, the decoding unit 1730 decodes the first determination bit value using the second determination bit value.

In this case, steps S1910 to S1911 have been described assuming a bitstream is generated by the encoding method described with reference to FIG. 13. If a bitstream is generated by the encoding method described with reference to FIG. 12, it will be apparent to those skilled in the art that steps S1910 to S1911 may be omitted.

In step S1920, the decoder 1730 extracts a sign symbol from the bitstream to be decoded, and in step S1921, the decoder 1730 extracts a sign bit constituting a sign symbol that is a unit of a predetermined sign bit.

In this case, although the decoding unit 1730 interprets the sign symbols as a bundle of sign bits in step S1921, it will be apparent to those skilled in the art that other components of the decoding apparatus may perform this.

In operation S1930, the prediction unit 1740 calculates a prediction value for the target pixel to be reconstructed. The prediction process is the same as the prediction process at the time of encoding described with reference to FIGS. 7 to 9. However, since the prediction process at the time of encoding is for a pre-stored image, the prediction process at decoding requires reconstruction of each pixel value, and thus the pixel value reconstruction and prediction process should be repeated one pixel value. .

In operation S1940, the reconstruction unit 1750 reconstructs the original pixel value by using Equation 3 with the prediction value, the error value, and the sine bit.

[Equation 3]

이면,

If,

는 에러값,

는 대상 화소의 화소값,

는 대상 화소의 예측값,

는 화소값 및 예측값의 비트수,

는 사인비트에 따라 1 또는 -1을 가지는 사인값,

는 미리 정해진 최대 에러값이다.

Is the error value,

Is the pixel value of the target pixel,

Is the predicted value of the target pixel,

Is the number of bits of the pixel value and the predicted value,

Is a sine value with 1 or -1 depending on the sine bit,

Is a predetermined maximum error value.

In operation S1950, the restoration unit 1750 restores the original pixel value, and then determines whether the restored pixel is the last pixel of the image.

If it is not the last pixel, steps S1930 to S1940 are performed to calculate a prediction value for the next pixel using the pixels reconstructed up to now.

If the last pixel, the decoding process is terminated.

It will be apparent to those skilled in the art that the method of the present invention as described above may be implemented as a program and stored in a computer-readable recording medium (CD-ROM, RAM, ROM, floppy disk, hard disk, magneto-optical disk, etc.).

1 is a block diagram of an encoding apparatus according to an embodiment of the present invention.

2 is a flowchart of an encoding method according to an embodiment of the present invention.

3 is a block diagram of a unit for explaining a variable block configuration;

4 is a diagram illustrating a quadtree generated when the input image of FIG. 3 is encoded according to an embodiment of the present invention. FIG.

FIG. 5 shows the division of a Lena image into blocks of size 64 × 64.

6 is a diagram for explaining a DPCM.

7 and 8 are diagrams for describing a method of predicting each pixel value of image data according to an embodiment of the present invention.

9 is a prediction algorithm considering an interface of image data according to an embodiment of the present invention.

10 is a diagram for explaining resetting prediction values according to an embodiment of the present invention;

11 is a diagram illustrating a distribution of a difference value between a target pixel value and a predicted value of image data according to an embodiment of the present invention.

12 is a flowchart illustrating a method of encoding by an encoder by using a decision bit value according to an embodiment of the present invention.

13 is a flowchart illustrating a method of encoding by an encoder by using a decision bit value according to another embodiment of the present invention.

14 is a flowchart illustrating a process of improving compression efficiency by additional code processing to a decision bit encoding method according to an embodiment of the present invention.

15 is an error value distribution diagram of a Lena image according to an embodiment of the present invention.

16 is a distribution diagram of decision bit values selected for each block in a Lena image according to an embodiment of the present invention.

17 is a block diagram illustrating a decision bit decoding apparatus according to an embodiment of the present invention.

18 is a flowchart illustrating a decision bit decoding method according to an embodiment of the present invention.

19 is a flowchart illustrating a method of decoding a bitstream by a decoding unit and a reconstructing unit according to an embodiment of the present invention.

Claims (41)

In the encoding device using the decision bit (Bit Precision), A prediction unit to determine a prediction method of each target pixel of the image data to be encoded, and to calculate a prediction value of the target pixel by using the prediction method; An error value is calculated using the prediction value, the size of the block is sequentially increased from a size of a predetermined block by a predetermined unit, and a first decision bit value corresponding to the error value corresponding to the block is selected. An encoder which encodes the error value; And A variable block configuration unit configured to calculate variable block configuration information indicating a block configuration in which the compression capacity is minimized using the encoding result of the block, The encoder generates a bit stream by encoding the block according to the calculated variable block configuration information, and inserts one or more of the variable block configuration information and the first decision bit value into the bit stream. The first decision bit value is a value having a minimum compression capacity when the error values are respectively encoded using decision bit values 0 to n, and n is a natural number. The method of claim 1, The encoder further calculates a sign bit using the prediction value, At least one of the variable block configuration information, the sign bit, and the decision bit value is inserted into the bitstream, And the sine bit is a sign of a difference between a pixel value of the target pixel and the prediction value. The method of claim 1, And the prediction unit determines the prediction method based on whether or not an upper boundary of the target pixel includes an interface. The method of claim 2, And the predicted value is a value calculated by a predetermined method that is equal to or similar to the target pixel value by using a value of a pixel adjacent to the target pixel. The method of claim 4, wherein The predicted value

Is calculated using the following equation, wherein T is a preset threshold value, A and B are the left and upper pixel values of the target pixel, and C, D, E, F and G are pixels corresponding to B. A coding apparatus, characterized in that the pixel values are left, right, top left, top, and top right as reference. [Equation]

If is

ego,

If is

,

,

And

Of the minimum

If is

ego,

If is

ego,

If is

ego,

If is

being. The method of claim 5, The error value and the sine bit are calculated using the following equation,

Is the error value of the target pixel,

Is a pixel value of the target pixel,

Is the predicted value of the target pixel,

Is the sign bit,

Is characterized in that the predetermined maximum error value, And the error value is a magnitude of a difference between a pixel value of the target pixel and the prediction value.  [Equation]

ego,

Back side

The method of claim 1, The pre-specified block size is 2 ⁿ X2 ⁿ coding apparatus characterized in that (n is a natural number) provided that, n is sequentially increased by one. The method of claim 2, And the encoding unit interprets the sine bit into a sine symbol that is a predetermined bundle unit of the sine bit. The method of claim 1, The encoder selects a second decision bit value corresponding to the first decision bit value, and further encodes the first decision bit value using the second decision bit value. And wherein the second decision bit value is a value having a minimum compression capacity obtained by encoding the first decision bit value using decision bit values 0 to n, and n is a natural number. The method of claim 1, If the frequency of the error value 0 among the error values is less than the frequency of the error value 2 ^k -1 (k is the first decision bit value), the coder encodes the sign of the error value 0 and the error value 2 ^k -1. Exchange it, And a bit indicating whether the code of the error value 0 is exchanged into the bitstream. The method of claim 9, The encoder compares the frequency of the first decision bit value 1 and the frequency of the remaining first decision bit value among the first decision bit values, Modifying the sign of the first decision bit value 1 to the sign of the first decision bit value t + 1 (where t is the decision bit value which is the largest of the decision bit values having a frequency greater than the decision bit value 1), The sign of the first decision bit value 2 to t is modified to the sign of the first decision bit value 1 to t-1, And a predetermined 3 bits for distinguishing the modified code of the first decision bit value 1 is inserted into the bitstream. In the decoding device using the bit (Bit Precision), A block generation unit generating a variable block by using variable block configuration information extracted from a bit stream to be decoded; A decoder configured to decode the bitstream by using the first decision bit value extracted from the bitstream and the variable block configuration information to calculate an error value; And A prediction unit to determine a prediction method of each target pixel of the image data to be decoded, and to calculate a prediction value of the target pixel using the prediction method; And a reconstruction unit configured to calculate a pixel value using the error value and the predicted value. The method of claim 12, And the predicted value means a value calculated by a predetermined method that is equal to or similar to the target pixel value by using values of pixels adjacent to the target pixel. The method of claim 12, And the decoding unit decodes the first decision bit value encoded using the second decision bit value extracted from the bitstream. The method of claim 14, And the decoding unit decodes the error value encoded using the first decision bit value. The method of claim 14, The decoding unit extracts a bit indicating a modified position of the first decision bit value 1 among the first decision bit values during encoding in the bitstream, and converts a sign of the first decision bit value 1 into a first decision bit t ( And a sign of the first decision bit value 2 to t to a sign of the corresponding first decision bit value 1 to t-1. The method of claim 15, The decoding unit extracts a bit indicating whether the code of the error value 0 during the encoding in the bitstream is exchanged, and if the code is exchanged, the code of the error value 0 and the error value 2 ^k -1 (k Replaces the code of the first decision bit value). The method of claim 13, The predicted value

Is calculated using the following equation, wherein T is a preset threshold value, A and B are the left and upper pixel values of the target pixel, and C, D, E, F and G are pixels corresponding to B. A decoding apparatus, characterized in that the pixel values are left, right, upper left, upper, and upper right as reference. [Equation]

If is

ego,

If is

,

,

And

Of the minimum

If is

ego,

If is

ego,

If is

ego,

If is

being. The method of claim 12, The pixel value is calculated using the following equation,

Is the error value of the target pixel,

Is a pixel value of the target pixel,

Is the predicted value of the target pixel,

Is a sine bit extracted from the bitstream,

Is characterized in that the predetermined maximum error value, And the error value and the sine bit are a magnitude and a sign of a difference between a pixel value of the target pixel and the prediction value. [Equation]

If,

In the encoding method using the decision bit (Bit Precision), (a) determining a prediction method corresponding to each target pixel of the image data to be encoded; (b) calculating a predicted value of the target pixel corresponding to the prediction method; (c) calculating an error value using the predicted value, sequentially increasing the size of the block by a predetermined unit from a size of a predetermined block, and selecting a decision bit value corresponding to the error value corresponding to the block. Making; (d) encoding the error value using the predetermined decision bit value; (e) calculating variable block configuration information for minimizing a compression capacity by using the encoded result of the block; And (f) encoding a block according to the variable block configuration information to generate a bit stream, One or more of the variable block configuration information and the decision bit value are inserted into the bitstream, and the predetermined decision bit value is a value having a minimum compression capacity in which the error values are encoded using decision bit values 0 to n, respectively. And n is a natural number. The method of claim 20, Step (c) further calculates a sign bit using the prediction value, At least one of the variable block configuration information, the sign bit, and the decision bit value is inserted into the bitstream, And the sine bit is a sign of a difference between a pixel value of the target pixel and the prediction value. The method of claim 20, In step (a), the prediction method is determined based on whether or not an upper boundary of the target pixel includes an interface. The method of claim 20, And the predicted value means a value calculated by a predetermined method that is equal to or similar to the target pixel value by using values of pixels adjacent to the target pixel. The method of claim 23, wherein The predicted value

Is calculated using the following equation, wherein T is a preset threshold value, A and B are the left and upper pixel values of the target pixel, and C, D, E, F and G are pixels corresponding to B. A coding method, characterized in that the pixel values are left, right, top left, top, and top right respectively. [Equation]

If is

ego,

If is

,

,

And

Of the minimum

If is

ego,

If is

ego,

If is

ego,

If is

being. The method of claim 21, The error value and the sine bit are calculated using the following equation,

Is the error value of the target pixel,

Is a pixel value of the target pixel,

Is the predicted value of the target pixel,

Is a sine bit,

Is characterized in that the predetermined maximum error value, And the error value is a magnitude of a difference between a pixel value of the target pixel and the prediction value. [Equation]

ego,

Back side

The method of claim 20, The pre-specified block size is 2 ⁿ X2 ⁿ coding method characterized in that (n is a natural number) provided that, n is sequentially increased by one. The method of claim 21, And the sign bit is interpreted as a sign symbol which is a predetermined bundle unit of the sign bit. The method of claim 20, The step (c) includes (g1) selecting a first decision bit value corresponding to the error value. The method of claim 28, Step (c) further includes (g2) selecting a second decision bit value corresponding to the first decision bit value after step (g1), And the second decision bit value is a value having a minimum compression capacity obtained by encoding the first decision bit value using decision bit values 0 to n, and n is a natural number. The method of claim 28, In the step (d), if the frequency of the error value 0 is smaller than the frequency of the error value 2 ^k -1 (k is the first decision bit value), the sign of the error value 0 and the error value k are determined. Exchange it, And inserting a bit indicating whether or not the sign of the error value 0 is exchanged into the bit stream. The method of claim 28, In step (d), the frequency of the first decision bit value 1 among the first decision bit values is compared with the frequency of the remaining first decision bit values. Modifying the sign of the first decision bit value 1 to the sign of the first decision bit value t + 1 (where t is the decision bit value which is the largest of the decision bit values having a frequency greater than the decision bit value 1), The sign of the first decision bit value 2 to t is modified to the sign of the first decision bit value 1 to t-1, And a predetermined three bits for distinguishing the modified code of the first decision bit value 1 are inserted into the bitstream. In the decoding method using bit precision, (a) extracting variable block configuration information from a bit stream to be decoded to generate a variable block; (b) extracting a decision bit value from the bitstream; (c) decoding the bitstream using the decision bit value and the variable block to calculate an error value; (d) determining a prediction method of each target pixel of the image data to be decoded, and calculating a prediction value of the target pixel using the prediction method; and (e) restoring image data using the error value and the prediction value. The method of claim 32, And the predicted value means a value calculated by a predetermined method that is equal to or similar to the target pixel value by using a value of a pixel adjacent to the target pixel. The method of claim 32, Step (b) includes (e1) extracting a first decision bit value corresponding to the error value from the bitstream, Step (c) includes (f1) decoding the error value encoded using the first decision bit value. The method of claim 34, Step (b) further includes extracting a second decision bit value from the bitstream before step (e1), And the step (c) further includes decoding the first decision bit value encoded using the second decision bit value before step (f1). The method of claim 34, Step (c) extracts a bit indicating whether or not the code of error value 0 is exchanged during encoding in the bitstream, and if the code is indicated to be exchanged, the code of error value 0 and error value 2 among the error values. ^{and k-} 1 (where k is the first decision bit value). The method of claim 34, Step (c) extracts a bit indicating a modified position of a first decision bit value 1 of the first decision bit values during encoding in the bitstream, and first determines a sign of the first decision bit value 1. And replaces the sign of the first decision bit value 2 to t with the sign of the first decision bit value 1 to t-1 corresponding to the sign of the bit value t (the modified position). The method of claim 33, wherein The predicted value

Is calculated using the following equation, wherein T is a preset threshold value, A and B are the left and upper pixel values of the target pixel, and C, D, E, F and G are pixels corresponding to B. The decoding method, characterized in that the pixel values of the left, right, upper left, upper, and upper right respectively. [Equation]

If is

ego,

If is

,

,

And

Of the minimum

If is

ego,

If is

ego,

If is

ego,

If is

being. The method of claim 38, The pixel value is calculated using the following equation,

Is the error value of the target pixel,

Is a pixel value of the target pixel,

Is the predicted value of the target pixel,

Is a sine bit extracted from the bitstream,

Is characterized in that the predetermined maximum error value, And the error value and the sine bit are the magnitude and sign of the difference between the pixel value of the target pixel and the prediction value. [Equation]

If,

In the recording medium on which a program of instructions that can be executed by an encoding apparatus is tangibly implemented to perform an encoding method using bit precision, and which records a program that can be read by the encoding apparatus. (a) determining a prediction method corresponding to each target pixel of the image data to be encoded; (b) calculating a predicted value of the target pixel corresponding to the prediction method; (c) calculating an error value using the predicted value, sequentially increasing the size of the block by a predetermined unit from a size of a predetermined block, and selecting a decision bit value corresponding to the error value corresponding to the block. Making; (d) encoding the error value using the predetermined decision bit value; (e) calculating variable block configuration information for minimizing a compression capacity by using the encoded result of the block; And (f) generating a bit stream by encoding the block according to the variable block configuration information, At least one of the variable block configuration information and the determination bit value is inserted into the bitstream, Wherein the selected decision bit value is a value having a minimum compression capacity in which the error values are encoded using decision bit values 0 to n, and n is a natural number. In the recording medium in which a program of instructions that can be executed by a decoding apparatus is tangibly implemented to perform a decoding method using bit precision, and in which a program which can be read by the decoding apparatus is recorded, (a) extracting variable block configuration information from a bit stream to be decoded to generate a variable block; (b) extracting a decision bit value from the bitstream; (c) decoding the bitstream using the decision bit value and the variable block to calculate an error value; (d) determining a prediction method of each target pixel of the image data to be decoded, and calculating a prediction value of the target pixel using the prediction method; (e) a recording medium having recorded thereon a program for performing restoration of the image data using the error value and the prediction value.

Images