JP4331992B2 - Image encoding method, image decoding method, image encoding device, image decoding device, program thereof, and program recording medium thereof - Google Patents

Description

  The present invention relates to an image encoding / decoding technique, and in particular, an approximate image of a second image can be obtained from a part of difference information between the first image and the second image and the first image information. The present invention relates to an image encoding device, an image decoding device, an image encoding method, an image decoding method, a program thereof, and a program recording medium thereof.

  There is a scalable encoding method in which an image is irreversibly encoded to generate base layer encoded data, and difference information between the decoded image and the original image is encoded to generate enhancement layer encoded data.

  For example, in the MPEG-4 Simple Scalable encoding tool, base layer encoded data is generated by irreversibly encoding an original image at a preset encoding rate, and difference information between the decoded image and the original image is calculated between frames. Enhancement layer encoded data is created as a prediction residual in encoding.

  The base layer and enhancement layer employ a method of quantizing image information or prediction residual by performing discrete cosine transform (DCT), which is a kind of orthogonal transform, and controlling the quantization parameter to achieve the desired coding. Realize the rate.

Note that, in enhancement layer coding, a method of directly coding a difference value without orthogonally transforming difference information is described in Non-Patent Document 1 below (see Non-Patent Document 1).
Kizen, Uekura, Yashima, "A Study on Pixel Area Scalable Coding", The Institute of Image Electronics Engineers, 202nd Research Meeting, 2003

  High-frequency components found in the edge of image information are important for subjective quality. However, such high frequency components cannot obtain high coding efficiency by orthogonal transform. In conventional scalable coding such as MPEG-4 Simple Scalable, low frequency components are mainly encoded in the base layer and high frequency components are encoded in the extension layer. However, in the method using orthogonal transformation, the high frequency components are efficiently used. Since encoding is not possible, there is a problem that the amount of encoded data in the enhancement layer increases.

  In addition, according to the method described in Non-Patent Document 1, it is possible to appropriately control the position where high-frequency components are encoded, so that the subjective quality near the edge is improved as compared with the method using orthogonal transformation. Is possible. However, there is a problem in that high-frequency components cannot originally obtain high coding efficiency because the correlation between adjacent pixels is low.

  The present invention solves the above-mentioned problems of the prior art, and reduces the amount of encoded data when obtaining the second image by encoding / decoding difference image information between the first image and the second image. It is an object of the present invention to provide an image encoding device, an image decoding device, an image encoding method, an image decoding method, a program thereof, and a program recording medium thereof.

In order to solve the above problems, the first invention provides an image of image B from image information of image A when image A used as a reference image at the time of encoding and image B to be encoded exist. An image encoding method for encoding an image B by encoding an information difference, wherein a difference creating step for obtaining an absolute value of difference information at each pixel position and creating a bit string by binary display, and a difference creating step In the bit string created in step 1, the bit truncation step in which all the values of the lower bits than the most significant bit having a value of 1 are all zero, and the bit string created in the bit truncation step are preset from the upper level to the lower level. and with the number of bits into a plurality of planes, the bit extracting a bit string of each plane, each plane, extract a bit extraction step A bit encoding step for encoding the most significant bit position information of the bit sequence and a positive / negative encoding step for encoding positive / negative information indicating whether the bit sequence encoded in the bit encoding step is positive or negative It is characterized by that.

The second invention is to encode a difference between image information of image B from image information of image A when image A used as a reference image at the time of encoding and image B to be encoded exist. The image encoding method for encoding the image B using the difference creation step for obtaining the absolute value of the difference information at each pixel position and creating a bit string by binary display, and the bit string created by the difference creation step, A bit truncation step in which the values of all the bits lower than the most significant bit, which is 1, are all zero, and a bit string created by the bit truncation step is divided into a plurality of planes with a preset number of bits from the upper level to the lower level. separated into a bit extracting a bit string of each plane, each plane, the top of bi bit strings extracted by the bit extraction step And DOO position information, said bit sequence and executes a bit encoding step comprising positive and negative encoding positive and negative information indicating whether positive or negative, the.

  The third invention is the image coding method according to the first or second invention, characterized in that the bit extraction step and the bit coding step are executed for each region into which the image B is divided.

  A fourth invention is an image coding method according to any one of the first to third inventions, wherein in the bit coding step, a probability table used for arithmetic coding using absolute values of difference information at surrounding pixel positions is provided. It is changed and arithmetic coding is performed.

The fifth invention is an image decoding method for decoding an image B when there is image information of the image A used as a reference image at the time of decoding and encoded data of the image B encoded by the first invention. In addition, for each plane, a positive / negative decoding step for decoding positive / negative information indicating whether the encoded data of the difference image information is positive or negative, and for each plane, from the encoded data of the difference image information, A bit decoding step for decoding the bit position information in the plane and setting the bit at the bit position obtained by decoding in the plane to 1 to obtain a fixed-length bit string, and the bit string obtained in the bit decoding step as a difference between the pixel positions A bit setting step for setting the bit string position of the corresponding plane in the absolute value image, and creating difference image information from the difference absolute value image and positive / negative information of the pixel position A difference generating step, by adding the image information and the difference image information of image A that is stored as a reference image in advance reference image memory, and executes an image generation step of generating image information.

The sixth invention is an image decoding method for decoding an image B when there is image information of the image A used as a reference image at the time of decoding and encoded data of the image B encoded by the second invention. For each plane, the most significant bit position information in the plane is decoded from the encoded data of the difference image information, and a bit string in which the bit at the bit position obtained by decoding in the plane is set to 1, A bit decoding step including positive / negative that decodes positive / negative information indicating whether it is negative or negative, and a bit setting step for setting the bit string obtained in the bit decoding step to the bit string position of the corresponding plane in the difference absolute value image at the pixel position; a difference generating step of generating a difference image information from the difference absolute value image and the positive and negative information of the pixel position is stored as the reference image in advance reference image memory By adding image information and the difference image information of image A are, and executes an image generation step of generating image information.

A seventh invention is the image decoding method according to the fifth or sixth invention, wherein the difference absolute value image set at the bit setting step and the value is 1 before executing the difference creation step. A bit addition step is performed in which the value of a bit lower than the first bit is set to a predetermined value .

The eighth invention is the image decoding method according to the fifth or sixth invention, wherein the difference absolute value image set in the bit setting step before the difference creation step is executed and the highest value is 1. A bit addition step is executed to set a value of a bit lower than the bit of 1 to a value specified by a parameter indicating a calculation method of a value of a lower bit included in encoded data .

  The ninth invention is the image decoding method of the fifth to eighth inventions, characterized in that a bit setting step and a bit decoding step are executed for each region into which the image B is divided.

  The tenth invention is the image decoding method according to any of the fifth to ninth inventions, wherein in the bit decoding step, the probability table used for arithmetic decoding is changed using the absolute value of the difference information at the surrounding pixel positions, It is characterized by performing arithmetic decoding.

  According to the image coding method of the first invention and the image decoding method of the fifth or seventh invention, an image is obtained from a part of the difference image information of the image B with respect to the image A and the image information of the image A on the decoding side. When creating encoded data of an image B that can obtain image information close to B, the amount of encoded data can be reduced as compared with encoding differential image information.

  In the present invention, not all the difference image information is encoded, but information other than the most significant bit having a value of 1 is set to 0 and encoded. Therefore, in each pixel, the most significant bit position having a value of 1 can be an encoding target. Hereinafter, the most significant bit position having a value of 1 will be referred to as the most significant bit position.

  For example, when the difference image information is represented by 8 bits, conventionally, it has been necessary to represent a value from 0 to 255 in each pixel. However, according to the present invention, the value may be from 0 to 8, If the value is fixed-length encoded, it may be 4 bits.

  Whether or not there is 1 at the least significant bit position in the difference image information is that the amount of change in the value in the difference image information is 1, and is not a significant difference subjectively. Therefore, if it is not necessary to distinguish whether the bit value of the least significant bit position is 0 or 1, the most significant bit position having a value of 1 that is the object of encoding in the present invention is It is sufficient to assume 0 to 7, and 3 bits are sufficient when the value is fixed-length encoded. Such difference image information is independently encoded positively or negatively.

  On the other hand, at each pixel, the value 0 is decoded at the decoding side except for the most significant bit position. The value 0 may be used as the difference image information as it is, or may be set to a value obtained according to a preset condition. For example, all lower bits below the most significant bit position may be set to 1. Or you may set to the half value of the value when all the lower bits below the most significant bit position are set to 1.

  Furthermore, in the present invention, the pixel position having a large difference and the pixel position having a small difference may be separated and encoded. According to the present invention, since the difference image information is encoded only once for each pixel, if the pixel position that is not the object of encoding is set to a value of 0, only the number of planes to be separated depending on the magnitude of the difference is set. Binary or multivalued rectangular image information can be constructed. This may be the encoding target.

  For example, when the absolute value of the difference information is 100 in decimal notation, it is 01100100 in binary notation, but when the value other than the most significant bit position is set to 0, it becomes 01000000. When this is divided into 4 bits and separated into 2 planes, 0100,0000 is obtained, so that it can be encoded as 5-plane (0-4) image information of 2 planes.

  The value in each plane indicates the most significant bit position. Since the upper level plane 0100 represents a large change in the difference, it corresponds to a lower frequency component. Therefore, for example, even if only the upper level plane is decoded on the decoding side, an image relatively close to the original image can be obtained subjectively.

  The number of planes to be divided and the bit depth (gradation) of each plane can be set arbitrarily if set in advance. In the case of a moving image, the number of planes and the bit depth of the plane may be changed for each frame. When the difference image information is encoded with one plane, the bit depth of the plane is the same as the bit depth of the image information.

  Note that the bit depth (gradation) may be changed for each plane. For example, 8-bit image information may be composed of 4 planes of 4 bits, 2 bits, 1 bit, and 1 bit from the upper level.

  In the present invention, the image sizes of the image A and the image B may be different. When the image A is smaller than the image B, the image information of the image A is up-sampled so as to have the same image size as the image B, and then the difference information is created. Conversely, if the image A is larger than the image B, the image information of the image A is down-sampled to the same image size as the image B, and then the difference information is created.

  For example, in encoding using the YCrCb color space, there is a method in which the image size of the Cr component and the Cb component is set to half the image size of the Y component. If the Cr component and the Cb component are up-sampled to the image encoded by this method and the encoding method of the present invention is applied, the image size of the Cr component and the Cb component is equal to the image size of the Y component. Can be generated.

  In the present invention, the bit depth (gradation) of the image information of the image A and the image B may be different. If the bit depth of the image A is different from the bit depth of the image B, the image information of the image A is converted to the same bit depth as that of the image B, and then difference information is created.

  For example, for image information in which each color component is represented by 8 bits in the YCrCb color space or RGB color space, it is possible to create encoded data that obtains image information in which each color component is represented by 10 bits or 12 bits. .

  As the binary or multi-level encoding method in each plane in the present invention, any encoding method may be used as long as it is lossless. For example, JBIG or JBIG is used as a binary encoding method, and JPEG-LS is used as a multi-level encoding method. In particular, in the case of multiple values, a lossless orthogonal transform or a subband division method may be used. This encoding method is preferably selected for each plane.

According to the image encoding method of the second invention and the image decoding method of the sixth or eighth invention, the most significant bit position and the positive / negative information in the absolute value of the difference image are encoded or decoded together. Can do. Since the difference is a positive, negative, or zero value, if the absolute value is N bits, at each pixel position,
-The positive most significant bit position is one of 1 to N,
-Or one of the negative most significant bit positions from 1 to N,
・ Or both positive and negative most significant bit positions are 0,
Either.

  At each pixel position, one of these (2N + 1) types of states is set. These (2N + 1) types of states may be associated with (2N + 1) numbers (hereinafter referred to as positive / negative bit positions) and may be used as encoding targets. As the positive and negative bit positions, for example, when the absolute value is 8 bits, when the positive most significant bit position is any one of 1 to 7, the numbers from 8 to 15 are associated with each other, and the negative most significant bit position is associated. A number from 6 to 0 is associated with each of the bit positions from 1 to 7, and a number of 7 is associated with both the positive and negative most significant bit positions being 0.

In addition, this method may be applied only to the upper M bits of which the absolute value is N bits. Also,
-The positive most significant bit position is one of 1 to M1,
-Or the negative most significant bit position is from 1 to M2,
・ Or both positive and negative most significant bit positions are 0,
As described above, the present invention may be applied to the range of the most significant bit position that is different between positive and negative.

  According to this method, since the information of the most significant bit position and the positive / negative information can be encoded together, the encoding efficiency can be further improved.

  According to the image encoding method of the third invention and the image decoding method of the ninth invention, the difference image information can be encoded for each region into which the screen is divided. As the region, a rectangular block such as a macro block is also suitable, and an arbitrarily shaped region is also suitable.

  This improves the subjective quality by encoding high-frequency components only in specific areas, or encodes high-frequency components in all areas and selects areas that are improved in quality according to the transmission rate, and selects only the selected areas. Decoding on the decoding side can improve the subjective quality. In particular, in the case of an arbitrary shape, a high frequency component can be encoded only for a specific image object or edge peripheral portion.

  As a method for specifying an area, the area division information may be explicitly specified by binary encoding or the like. An edge may be extracted and divided into regions on the decoding side. In this case, however, the edge extraction parameters must also be encoded.

  Alternatively, the entire screen may be encoded with the outside of the area regarded as an insignificant value. For example, if there is a screen consisting of two regions, region X and region Y, the entire screen with region X as a significant value, region Y as a non-significant value (for example, 0), and region X as a non-significant value (For example, 0), the entire screen with the region Y as a significant value is encoded. In this case, a plurality of planes of the same level exist for the same area. The number of planes to be divided and the bit depth (gradation) of each plane may be changed for each region.

  According to the image encoding method of the fourth invention and the image decoding method of the tenth invention, when encoding binary or multivalued information divided into planes, information on surrounding pixel positions is used. Arithmetic coding can be applied while changing the probability table. As a result, it is possible to adaptively apply arithmetic coding for each similar region, rather than simply performing arithmetic coding of each value, so that the coding efficiency can be improved.

  According to the present invention, encoded data of an image B is generated so that image information close to the image B can be obtained from a part of the difference image information of the image B with respect to the image A and the image information of the image A on the decoding side. In this case, the most significant bit position of the difference image information can be set as an encoding target, and the amount of encoded data can be reduced as compared to encoding the difference image information.

  It is also possible to encode the difference image information by separating it into a plurality of planes, so that the amount of encoded data to be decoded can be determined on the decoding side, and control to gradually decode high-frequency components is possible. It is also possible to do.

  Embodiments of the present invention will be described with reference to the drawings. A method of encoding difference information between an image A of 320 pixels by 240 pixels vertically expressed by 8 bits and an image B of 640 pixels by 480 pixels expressed by 10 bits will be described. On the decoding side, a method of obtaining an approximate image of the image B from the image A and the encoded data is shown. It is assumed that the difference information is encoded in a plane every 2 bits.

  First, the image encoding device will be described. FIG. 1 shows an outline of the apparatus. The image encoding device 1 generates an image B input unit 102 that inputs an image B of an encoding target image, an image A input unit 101 that inputs an image A of a reference image, and difference image information between the images A and B The difference creation unit 103, the bit truncation unit 107 that sets the value other than the most significant bit position of the absolute value of the difference information at each pixel position to 0, and the value created by the bit truncation unit 107, the difference between positive and negative A differential accumulation memory 104 that separately accumulates the values, a bit extraction unit 105 that extracts a value from every two bits from the top, a bit encoding unit 106 that encodes the two bits extracted by the bit extraction unit 105, , And a positive / negative encoding unit 108 that encodes positive / negative information.

  In creating the difference image information, the difference creating unit 103 enlarges the image information of the image A from 8 bits to 10 bits and doubles the image size. First, the image size is doubled and then the gradation is increased.

Linear interpolation between adjacent pixels is used to enlarge to 2 times. A schematic diagram of linear interpolation is shown in FIG. “>> 1” represents a 1-bit shift to the right. Further, it is assumed that the gradation is increased according to the equation (1). I indicates image information of the image A, and O indicates a value changed to 10 bits.

O = I × 4 (1)
The difference creating unit 103 calculates the difference from the image B after the image A is made 10 bits in this way and has an image size of 640 × 480 in length. In addition, the difference accumulation memory 104 is prepared with a memory for a positive value and a negative value, and the bit truncation unit 107 is stored in a memory for a negative value at a pixel position where the difference is a positive value. The value 0 is accumulated, and the value 0 is accumulated in the positive value memory at the pixel position where the difference is a negative value.

  The bit encoding unit 106 encodes 2 bits in the raster scan direction from the upper left pixel. As the 2-bit value, the most significant bit position is the encoding target. Since binary representation can take any of 00, 01, and 10, this value is gray-coded and encoded using JBIG gray-scale coding.

  Further, when the bit encoding unit 106 encodes image information for a positive value, the positive / negative encoding unit 108 sets a value 0 to the positive / negative information, and the bit encoding unit 106 sets an image for a negative value. When encoding information, the value 1 is set in the positive / negative information, and the positive / negative information is 1-bit fixed-length encoded in each plane.

  Under this assumption, the image information of the image B is encoded as follows. First, the image A input unit 101 and the image B input unit 102 each input an image. The difference creation unit 103 creates difference image information from these pieces of image information.

  The bit truncation unit 107 sets the absolute value of the difference information other than the most significant bit position to 0, and if the difference is a positive value, accumulates it in the difference accumulation memory 104 for a positive value. Is stored in the difference accumulation memory 104 for negative values. The bit truncation unit 107 accumulates a value 0 in the negative value memory at the pixel position where the difference is positive, and stores a value 0 in the positive value memory at the pixel position where the difference is negative. accumulate.

  Subsequently, the difference information accumulated in the difference accumulation memory 104 for positive values is encoded from the upper level 2 bits to the lower level for each plane as follows. First, the positive / negative encoding unit 108 encodes positive / negative information indicating that the plane is encoded data for a positive value.

  Next, the first level first plane is encoded from the upper left pixel as follows. The bit extraction unit 105 extracts 2 bits included in the plane for the difference information of the pixel position. The bit encoding unit 106 encodes a 2-bit value. This operation is repeated until the lower right pixel position is encoded.

  Subsequently, the second plane as the next higher level is encoded from the upper left pixel according to the same procedure as described above. Since 10-bit image information is encoded with a 2-bit plane, a maximum of 5 planes can be encoded.

  After all the difference information stored in the positive value difference storage memory 104 is encoded, the difference information stored in the negative value difference storage memory 104 is encoded in the same manner as described above.

  Next, the decoding device will be described. FIG. 3 shows an outline of the apparatus. The image decoding apparatus 2 is obtained by the positive / negative decoding unit 207 that decodes the positive / negative information of the decoding target plane, the bit decoding unit 201 that decodes a 2-bit fixed length bit string from the encoded data of the decoding target plane, and the bit decoding unit 201 A bit setting unit 202 that sets 2 bits as appropriate bits of a difference absolute value image, a reference image memory 205 that stores image information of image A, and difference generation that generates difference image information from the difference absolute value image and positive / negative information Image creation for creating image information of image B by adding unit 206, difference accumulation memory 203 that accumulates difference image information separately for positive and negative, and image information of image A and difference image information The unit 204 is provided.

  In creating the decoded image, the image creating unit 204 enlarges the image information of the image A from 8 bits to 10 bits and doubles the image size. First, the image size is doubled and then the gradation is increased.

  Linear interpolation between adjacent pixels is used to enlarge to 2 times. A schematic diagram of linear interpolation is shown in FIG. Further, it is assumed that the gradation is increased according to the above formula (1). The bit decoding unit decodes the encoded data using JBIG grayscale decoding, decodes gray, and decodes 00, 01, and 10 in binary representation, and sets the obtained value to the most significant bit position. To do.

  Under such a premise, the encoded data is decoded as follows. First, the positive / negative decoding unit 207 decodes the positive / negative information of the plane. Next, each plane is decoded from the upper plane to the lower plane as follows. For the upper left pixel, the bit decoding unit 201 decodes a 2-bit value from the encoded data of the difference image information.

  The bit setting unit 202 sets the decoded value in the difference absolute value image. The difference creating unit 206 accumulates the difference absolute value image in the positive value memory or the negative value memory of the difference accumulation memory 203 according to the positive / negative information.

  The above operation is repeated until the lower right pixel position is decoded. Subsequently, the second plane as the next higher level is decoded from the upper left pixel according to the same procedure as described above. The above processing is performed on the number of planes input to the image decoding device 2.

  After decoding all the planes, the image creation unit 204 creates a decoded image from the image information of the image A stored in the reference image memory 205 and the difference image information stored in the difference storage memory 203.

  In this embodiment, all the planes are decoded. However, the procedure for creating the difference image information from each plane may be completed up to the intermediate plane, and the decoded image may be created. Only encoded data of either a positive value or a negative value may be decoded.

  In this embodiment, 10-bit difference image information is encoded with a total of 10 planes with 5 positive and negative planes, but may be encoded with 2 planes with 1 positive and negative planes. The number of bits in each plane is not limited to 2 bits.

  It is also possible to encode / decode only a specific area, or to select an area to be decoded. In this case, the difference information of the pixel positions other than the specific area is set to 0, and encoding / decoding is performed as a separate plane for each area.

  Alternatively, the value of the lower bit below the most significant bit position may be obtained according to a preset condition. For example, all lower bits below the most significant bit position may be set to 1. Alternatively, it may be set to a value half that of a value when all the lower bits below the most significant bit position are set to 1.

  Further, this condition may be changed depending on the plane. For example, if the most significant bit position is included in the most significant plane, set the value to one half of the value when all the lower bits below the most significant bit position are set to 1, and the most significant If the bit position is not included in the most significant plane, all the lower bits may be set to 0.

  In this way, it is also preferable to specify the method for calculating the value of the lower bits below the most significant bit position (referred to as bit reproduction information) for each plane using parameters. FIG. 4 shows a configuration diagram of an image encoding device when the bit reproduction information is designated by a parameter, and FIG. 5 shows a configuration diagram of the image decoding device.

  In addition to the configuration of the image encoding device 1 shown in FIG. 1, the image encoding device 3 includes a bit reproduction information encoding unit 301 that encodes bit reproduction information. The image decoding device 4 includes the image shown in FIG. In addition to the configuration of the decoding device 2, the bit reproduction information decoding unit 402 for decoding the bit reproduction information and the value of the lower bit below the most significant bit position in the difference absolute value image set by the bit setting unit 202 are changed. A bit adder 401.

  For example, the bit reproduction information encoding unit 301 of the image encoding device 3 shown in FIG. 4 encodes bit reproduction information indicating a method for calculating the value of the lower bit below the most significant bit position for each plane.

  In addition, the bit reproduction information decoding unit 402 of the image decoding device 4 shown in FIG. 5 decodes the bit reproduction information, and the bit addition unit 401 is a difference absolute value image set by the bit setting unit 202 and is not more than the most significant bit position. The difference creation unit 206 stores the difference absolute value image in the positive value memory or the negative value memory of the difference storage memory 203 according to the positive / negative information.

  After decoding all the planes, the image creation unit 204 creates a decoded image from the image information of the image A stored in the reference image memory 205 and the difference image information stored in the difference storage memory 203.

  In the image encoding device 1 shown in FIG. 1 and the image encoding device 3 shown in FIG. 3, the bit sequence extracted by the bit extraction unit 105 is encoded by the bit encoding unit 106, and positive / negative indicating whether the bit sequence is positive or negative Although the information is encoded by the positive / negative encoding unit 108, the bit string extracted by the bit extraction unit 105 and the positive / negative information may be encoded together.

  In this case, in the image decoding device 2 shown in FIG. 3 and the image decoding device 4 shown in FIG. 5, instead of the positive / negative decoding unit 207 decoding positive / negative information, the bit decoding unit 201 decodes the positive / negative information together with the bit string. It will be.

  The above-described image encoding and image decoding processes can be realized by a computer and a software program. The program can be provided by being recorded on a computer-readable recording medium or via a network. is there.

  The present invention can be applied to the use of obtaining an approximate image of the second image from a part of difference information between the first image and the second image and the first image information.

It is a figure which shows the structural example of an image coding apparatus. It is a figure which shows the example of linear interpolation. It is a figure which shows the structural example of an image decoding apparatus. It is a figure which shows the structural example of an image coding apparatus. It is a figure which shows the structural example of an image decoding apparatus.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1,3 Image coding apparatus 2,4 Image decoding apparatus 101 Image A input part 102 Image B input part 103 Difference preparation part 104 Difference accumulation memory 105 Bit extraction part 106 Bit encoding part 107 Bit truncation part
DESCRIPTION OF SYMBOLS 108 Positive / negative encoding part 201 Bit decoding part 202 Bit setting part 203 Difference storage memory 204 Image creation part 205 Reference image memory 206 Difference creation part

Claims (20)

An image encoding method for encoding a second image by encoding a difference of image information of a second image from image information of the first image used as a reference image at the time of encoding. And
A difference creation step of obtaining an absolute value of difference information at each pixel position of the first image and the second image and creating a bit string by binary display;
A bit truncation step in which all the values of lower bits than the most significant bit having a value of 1 in the bit string created in the difference creation step are set to 0;
A bit extraction step of separating the bit string created in the bit truncation step into a plurality of planes with a preset number of bits from an upper level to a lower level, and extracting a bit string of each plane;
A bit encoding step for encoding the most significant bit position information of the bit string extracted in the bit extraction step for each plane;
A positive / negative encoding step for encoding positive / negative information indicating whether the bit string encoded in the bit encoding step is positive or negative;
An image encoding method comprising: An image encoding method for encoding a second image by encoding a difference of image information of a second image from image information of the first image used as a reference image at the time of encoding. And
A difference creation step of obtaining an absolute value of difference information at each pixel position of the first image and the second image and creating a bit string by binary display;
A bit truncation step in which all the values of lower bits than the most significant bit having a value of 1 in the bit string created in the difference creation step are set to 0;
A bit extraction step of separating the bit string created in the bit truncation step into a plurality of planes with a preset number of bits from an upper level to a lower level, and extracting a bit string of each plane;
A bit encoding step including positive and negative for encoding the most significant bit position information of the bit string extracted in the bit extraction step and positive / negative information indicating whether the bit string is positive or negative for each plane;
An image encoding method comprising: The image encoding method according to claim 1 or 2,
An image encoding method, wherein the bit extraction step and the bit encoding step are executed for each region obtained by dividing the second image. In the image encoding method according to claim 1, claim 2 or claim 3,
An image encoding method, wherein in the bit encoding step, arithmetic coding is performed by changing a probability table used for arithmetic encoding using an absolute value of difference information at surrounding pixel positions. The image decoding which inputs the encoded data of the 2nd image based on the difference with the image information of the said 1st image encoded by the image encoding method of Claim 1, and decodes the said 2nd image A method,
A positive / negative decoding step for decoding positive / negative information indicating whether the encoded data of the second image based on the difference is positive or negative for each plane;
For each of the planes, a bit string in which the most significant bit position information in the plane is decoded from the encoded data of the second image based on the difference, and the bit in the bit position obtained by decoding in the plane is set to 1. A bit decoding step to obtain
A bit setting step for setting the bit string obtained in the bit decoding step to the bit string position of the corresponding plane in the difference absolute value image at the pixel position;
A difference creating step for creating difference image information from the difference absolute value image of the pixel position and the positive / negative information;
By adding the image information of the first image stored in the reference image memory that stores the image information of the first image used in advance as the reference image and the created difference image information, the second image An image creation step for creating image information of
The image decoding method characterized by performing. The image decoding which inputs the encoded data of the 2nd image based on the difference with the image information of the said 1st image encoded by the image encoding method of Claim 2, and decodes the said 2nd image A method,
For each of the planes, a bit string in which the most significant bit position information in the plane is decoded from the encoded data of the second image based on the difference, and the bit in the bit position obtained by decoding in the plane is set to 1. A bit decoding step including positive and negative information for decoding positive / negative information indicating positive or negative;
A bit setting step for setting the bit string obtained in the bit decoding step including the positive and negative to the bit string position of the corresponding plane in the difference absolute value image at the pixel position;
A difference creating step for creating difference image information from the difference absolute value image of the pixel position and the positive / negative information;
By adding the image information of the first image stored in the reference image memory that stores the image information of the first image used in advance as the reference image and the created difference image information, the second image An image creation step for creating image information of
The image decoding method characterized by performing. In the image decoding method according to claim 5 or 6,
A bit for setting a value of a bit lower than the most significant bit having a value of 1 to a predetermined value in the difference absolute value image set in the bit setting step before executing the difference creating step An image decoding method characterized by executing an addition step. In the image decoding method according to claim 5 or 6,
Before executing the difference creation step, in the difference absolute value image set in the bit setting step, the value of the lower bit than the most significant bit having a value of 1 is set to the lower bit included in the encoded data. An image decoding method comprising: performing a bit addition step of setting a value specified by a parameter indicating a value calculation method. In the image decoding method according to claim 5, claim 6, claim 7 or claim 8,
The image decoding method, wherein the bit setting step and the bit decoding step are executed for each region obtained by dividing the second image. In the image decoding method according to claim 5, claim 6, claim 7, claim 8 or claim 9,
An image decoding method, wherein, in the bit decoding step, arithmetic decoding is performed by changing a probability table used for arithmetic decoding using an absolute value of difference information at surrounding pixel positions. An image encoding device for encoding the second image by encoding a difference of image information of the second image from image information of the first image used as a reference image at the time of encoding. And
A difference creation unit for obtaining an absolute value of difference information at each pixel position and creating a bit string by binary display;
A bit truncation unit that sets all the values of lower bits than the most significant bit having a value of 1 in the bit string created by the difference creation unit;
A bit extraction unit that separates the bit sequence created by the bit truncation unit into a plurality of planes with a preset number of bits from the upper level to the lower level, and extracts the bit sequence of each plane;
A bit encoding unit that encodes the most significant bit position information of the bit string extracted by the bit extraction unit for each of the planes;
A positive / negative encoding unit that encodes positive / negative information indicating whether the bit string encoded by the bit encoding unit is positive or negative;
An image encoding device comprising: An image encoding device for encoding the second image by encoding a difference of image information of the second image from image information of the first image used as a reference image at the time of encoding. And
A difference creation unit for obtaining an absolute value of difference information at each pixel position and creating a bit string by binary display;
A bit truncation unit that sets all the values of lower bits than the most significant bit having a value of 1 in the bit string created by the difference creation unit;
A bit extraction unit that separates the bit sequence created by the bit truncation unit into a plurality of planes with a preset number of bits from the upper level to the lower level, and extracts the bit sequence of each plane;
A bit encoding unit including positive / negative for encoding the most significant bit position information of the bit string extracted by the bit extraction unit and positive / negative information indicating whether the bit string is positive or negative for each plane;
An image encoding device comprising: The image decoding which inputs the encoded data of the 2nd image based on the difference with the image information of the said 1st image encoded by the image encoding apparatus of Claim 11, and decodes the said 2nd image A device,
A reference image memory for storing image information of the first image used as a reference image in advance;
A positive / negative decoding unit that decodes positive / negative information indicating whether the encoded data of the second image based on the difference is positive or negative for each plane;
For each of the planes, a bit string in which the most significant bit position information in the plane is decoded from the encoded data of the second image based on the difference, and the bit in the bit position obtained by decoding in the plane is set to 1. A bit decoding unit for obtaining
A bit setting unit that sets the bit string obtained by the bit decoding unit to the bit string position of the corresponding plane in the difference absolute value image at the pixel position;
A difference creating unit that creates difference image information from the difference absolute value image of the pixel position and the positive / negative information;
An image creation unit for creating image information of the second image by adding the image information of the first image stored in the reference image memory and the created difference image information;
An image decoding apparatus comprising: The image decoding which inputs the encoded data of the 2nd image based on the difference with the image information of the said 1st image encoded by the image encoding apparatus of Claim 12, and decodes the said 2nd image A device,
A reference image memory for storing image information of the first image used as a reference image in advance;
For each of the planes, a bit string in which the most significant bit position information in the plane is decoded from the encoded data of the second image based on the difference, and the bit in the bit position obtained by decoding in the plane is set to 1. A bit decoding unit including positive and negative to decode positive / negative information indicating positive or negative,
A bit setting unit that sets the bit sequence obtained by the bit decoding unit including the positive and negative to the bit sequence position of the corresponding plane in the difference absolute value image at the pixel position;
A difference creating unit that creates difference image information from the difference absolute value image of the pixel position and the positive / negative information;
An image creation unit for creating image information of the second image by adding the image information of the first image stored in the reference image memory and the created difference image information;
An image decoding apparatus comprising: The image decoding device according to claim 13 or 14 ,
Prior to executing the processing by the difference creating unit, the value of the lower bit than the most significant bit having a value of 1 is set to a predetermined value in the difference absolute value image set by the bit setting unit. An image decoding apparatus comprising: a bit addition unit that performs The image decoding device according to claim 13 or 14 ,
Before executing the difference creating unit, in the difference absolute value image set by the bit setting unit, the value of the lower bit than the most significant bit having a value of 1 is set to the lower bit included in the encoded data. An image decoding apparatus comprising: a bit addition unit that sets a value specified by a parameter indicating a value calculation method.   An image encoding program for causing a computer to execute the image encoding method according to any one of claims 1 to 4.   The image decoding program for making a computer perform the image decoding method of any one of Claim 5-10.   An image encoding program recording medium in which a program for causing a computer to execute the image encoding method according to any one of claims 1 to 4 is recorded.   The image decoding program recording medium which recorded the program for making a computer perform the image decoding method of any one of Claim 5-10.

Images