JP4964084B2 - Image coding apparatus and image decoding apparatus - Google Patents

Description

  The present invention relates to an image encoding and image decoding technique used in an image processing apparatus, and in particular, an image composition apparatus that displays a base image through a superimposition image.

  In recent years, a method of displaying subtitle information or other image data on a photograph or video by superimposing the images on each other is generally performed. In order to perform the superimposition processing, a plurality of images are prepared, and one is used as a background image and the other is used as a superimposition image. Is displayed, and when the flag indicates non-transparent, the overlay image data is output, and when the flag indicates transparent, the background image data is output. There is a need.

  As one of the image encoding methods based on this superposition processing, a method disclosed in Japanese Patent Laid-Open No. 2005-45452 (Patent Document 1) is as follows. That is, when the image data is divided into a plurality of blocks and the reference level of the pixel value in each block is a predetermined value, the background data is selected for the data of the corresponding block as a portion corresponding to the data transmission process. . If the pixel value reference level of the block is any other value, the data of the corresponding block is decoded as it is to be output data.

JP-A-2005-45452

However, in the image processing apparatus described in Patent Document 1, since it is determined in block units whether or not to superimpose images, the overlay image is transmitted, and the place where the background image is displayed is specified in block units. Can only be done. For this reason, when a fine character or pattern is to be displayed superimposed on the superimposed image, there is a problem that the transmission process cannot be performed on the pixel that is originally subjected to the transmission process, and the image becomes unnatural.
In the present invention, when encoding is performed in units of blocks, a code is formed in consideration of pixels to be subjected to transmission processing, thereby enabling transmission processing in units of pixels.

An image encoding device according to the present invention is as follows.
Block dividing means for dividing the input image data of the overlay image into M × N (M and N are natural numbers) blocks, and signal level range calculating means for calculating the maximum and minimum pixel values in the block Encoding preparation means for calculating the reference level of the pixel of the block and the pixel value fluctuation range in the block from the maximum value and the minimum value of the obtained pixel value, and the maximum value and the minimum value of the pixel value obtained in the same manner Quantization threshold value calculation means for calculating a quantization threshold value of a pixel value for dividing the gradation level corresponding to the number of types of codes to be converted, and the gradation level of each pixel in the block as the quantization threshold value Depending on which of the plurality of ranges determined in accordance with each range, the gradation level of each pixel is replaced with a unique code corresponding to each range, and when a plurality of images are overlaid, transmission is performed. For the pixels that output the background image data, the quantization means for assigning a unique code of a different type from the above codes, the reference level of the pixels in the block, the pixel value fluctuation range, and the codes assigned to the individual pixels are summarized. An encoded data creating means for producing encoded data ,
The encoding preparation means calculates and outputs the reference level of the pixel of the block and the pixel value fluctuation range in the block, and outputs a signal indicating whether or not the block includes a pixel that transmits data in the block. Output included in the fluctuation range signal,
The quantization threshold value calculation means, when a pixel that transmits data is included in the block, a threshold value for dividing the gradation level into n stages, and when a pixel that transmits data is not included in the block, Each of the threshold values for dividing the gradation level into (n + 1) stages is obtained,
When the block includes pixels that transmit data, the quantization unit assigns a code indicating either a transparent pixel or a code indicating n gradation levels for each pixel, and assigns the data to the block. In the case where a transparent pixel is not included, a code indicating a gradation level of (n + 1) stages is assigned to each pixel.

An image encoding device according to the present invention is as follows.
The image data of the image for superimposition is divided and encoded for each M × N (M and N are natural numbers) blocks, and the reference level of the pixels in the block, the pixel value fluctuation range, and a plurality of images are superimposed and combined. An image decoding apparatus that inputs encoded data in which different types of codes assigned to pixels when a plurality of images are not overlapped with a pixel that is transmitted through and output the background image data and decodes the image data. There,
Gradation level calculation means for calculating the gradation level corresponding to the code of each pixel from the pixel reference level and pixel value fluctuation range of each block, and gradation level assignment means for replacing the code of each pixel with the gradation level Have
The gradation level calculation means calculates the gradation level corresponding to the code of each pixel from the pixel reference level and pixel value variation width of each block, and is included in the pixel value variation value width data in the block encoded data. It is determined whether the corresponding block includes a pixel that transmits data according to the flag, and if the corresponding block includes a pixel that transmits data, the gradation level corresponding to n types of codes is calculated and the corresponding block is calculated. Does not include pixels that transmit data, the gradation level corresponding to (n + 1) types of codes is calculated,
In the case where the block includes pixels that transmit data, if the code of the pixel indicates a transmission output, the gradation level assigning unit sets the gradation level of the pixel to the background of another image. In addition to replacing with the corresponding pixel gradation level , in other cases, the code of each pixel is replaced with the gradation level according to the correspondence between the n types of codes calculated by the gradation level calculating means and the gradation level, When the pixel that does not transmit data is not included, the code of each pixel is replaced with the gradation level according to the correspondence between the (n + 1) types of codes calculated by the gradation level calculation means and the gradation level .

According to the image encoding apparatus of the present invention, when assigning a code to each pixel which is divided into blocks, the pixel to be permeabilized to use a separate code indicating the pixels of the transmission process, in units of pixels It is possible to create encoded data corresponding to fine transparency processing.
In addition, since one of the types of codes used when the transparency processing is not considered is used as a code indicating the pixel corresponding to the transparency processing, the data length of the encoded data is changed even when the transparency processing is considered. It is possible to keep the same constant value as in the case of not considering.

  In addition, according to the image decoding apparatus of the present invention, when performing decoding, in order to refer to the background data corresponding to the individual code indicating the pixel on which the transparent processing is performed, the fine transparent processing is performed in units of pixels. It is possible to create decoded image data.

Embodiment 1 FIG.
FIG. 1 is a block diagram showing an image encoding apparatus according to Embodiment 1 of the present invention. In FIG. 1, 1 is a block dividing means for dividing image data into M × N (M and N are natural numbers) blocks, and 2 is a maximum value and a minimum value of gradation levels for pixels in the divided blocks. Signal level range calculating means for calculating, 3 an encoding preparation means for calculating a reference level and a pixel value fluctuation range of a pixel in the block from the maximum value and the minimum value of the gradation level, and 4 a reference level of the pixel Quantization threshold value calculation means for calculating a threshold value for quantization from the pixel value fluctuation range, 5 is a quantization means for assigning a quantized code corresponding to the gradation level of each pixel, and 6 is calculated. This is encoded data generating means for collectively combining the reference level of the pixels in the block, the pixel value fluctuation range, and the code assigned to each pixel into encoded data.

Next, the operation will be described. FIG. 2 is a flowchart showing an image processing method in the image coding apparatus according to the first embodiment. First, the image data input by the block dividing means 1 is divided into M × N (M and N are natural numbers) blocks (step ST1). Here, a case where each block is composed of 16 pixels of 4 horizontal pixels and 4 vertical pixels will be described.
Among the divided blocks, one block not subjected to the encoding process is selected (step ST2).

In the signal level range calculation means 2, the maximum value Lmax and the minimum value Lmin of the gradation level are obtained from the gradation level of each pixel of the selected block (step ST3).
The encoding preparation means 3 obtains the pixel reference level LA and the pixel value variation width LD of the block from the maximum value Lmax and minimum value Lmin of the pixel gradation level in the block and the distribution of the pixel gradation level in the block (step ST4).
A method for calculating the pixel reference level LA and the pixel value variation width LD will be described below.
First, threshold values P1 and P2 for determining reference pixels for calculating the pixel reference level LA and the pixel value fluctuation range LD are obtained by (Expression 1) and (Expression 2).
P1 = (Lmax + 3Lmin) / 4 (Formula 1)
P2 = (3Lmax + Lmin) / 4 (Formula 2)

  Here, the threshold values P1 and P2 are obtained as a value obtained by weighted average of Lmax and Lmin. At this time, the coefficient applied to the weighted average is a bit assigned in the gradation level distribution state of the image and the gradation level conversion of each pixel. It is possible to select a value that minimizes the quantization error according to the characteristics of the image of the system to which the present technology is applied. However, in order to perform processing assuming the same coefficient at the time of encoding and decoding at that time, in the same system, it is unified to a single coefficient, or a flag is provided in the encoded data and used It is necessary to have a mechanism to distinguish

Regarding the gradation level Xij (0 ≦ i, j <4) of the pixels included in the block,
Qmin is the average gradation level value of pixels satisfying Xij ≦ P1.
Qmax is an average value of gradation levels of pixels satisfying Xij> P2.
And

Using the values of Qmin and Qmax, the values of the pixel reference level LA and the pixel value variation width LD are obtained by the following (Expression 3) and (Expression 4).
LA = (Qmin + Qmax) / 2 (Formula 3)
LD = (Qmax−Qmin) (Formula 4)

  The quantization threshold calculation means 4 calculates a threshold for converting the gradation level of each pixel into a code (step ST5). The number of threshold values needs to correspond to the number of bits assigned to the code for converting the gradation level. Here, a case where the code assigned to one pixel is 2 bits will be described. When the code assigned to the pixel is 2 bits, four types of states can be assigned to each pixel. Of these, since one type is assigned to the state indicating the pixel to be subjected to the transmission processing, the remaining three types of codes corresponding to the gradation levels remain. Therefore, two values L1 and L2 are required as threshold values for dividing the gradation level into three ranges corresponding to the three types of codes.

The values of the thresholds L1 and L2 are based on the pixel reference level LA and the pixel value fluctuation width LD obtained in step ST4, and values of constant ratios of LA to LD are obtained as shown in (Expression 5) and (Expression 6) below. Find by adding or subtracting.
L1 = LA-LD / 4 (Formula 5)
L2 = LA + LD / 4 (Formula 6)

At this time, the ratio applied to the value of the pixel value fluctuation width LD is 1/4 here, but the quantization error depends on the gradation level distribution state of the image and the characteristics of the image of the system to which the present technology is applied. It is possible to select a value as small as possible. However, in order to perform processing assuming the same coefficient at the time of encoding and decoding at that time, in the same system, it is unified to a single coefficient, or a flag is provided in the encoded data and used It is necessary to have a mechanism to distinguish
FIG. 4 shows the encoding threshold distribution and the signal level distribution at the time of decoding when (Equation 5) and (Equation 6) are applied. FIG. 3 shows a signal level distribution in the conventional encoding method in which the gradation level of each pixel corresponds to four types of codes without considering the transparent pixels. In FIG. 4, the gradation level assigned to the conventional encoding method shown in FIG. 3 is reduced (four to three), so that the gradation reproducibility of the decoded image is lowered.

  If the code assigned to a pixel is 3 bits, eight types of states can be assigned to each pixel, and one of them is assigned to a state indicating a pixel to be subjected to a transmission process. There are seven remaining codes corresponding to. Therefore, six values L1 to L6 are required as threshold values for dividing the gradation level into seven ranges corresponding to the seven types of codes. An example of an arithmetic expression is shown in (Expression 7) to (Expression 12) below.

L1 = LA-LD × 3/8 (Expression 7)
L2 = LA-LD / 4 (Formula 8)
L3 = LA-LD / 8 (formula 9)
L4 = LA + LD / 8 (Formula 10)
L5 = LA + LD / 4 (Formula 11)
L6 = LA + LD × 3/8 (Formula 12)

One of the unprocessed pixels in the selected block is selected (step ST6).
The quantization means 5 confirms whether or not the selected pixel is a pixel to be subjected to transmission processing (step ST7). It is assumed that pixel data to be subjected to transparency processing in the input image is determined in advance together with the input image.
If the selected pixel is not a pixel to be subjected to the transmission process in step ST7, it is determined which of the areas formed by the threshold values L1 and L2 obtained in step ST5 corresponds to the gradation level of the selected pixel. Then, a code is assigned (step ST8). An example of an allocation method when the number of bits allocated to the code is 2 bits is shown below. The gradation level Xij of the selected pixel and the corresponding code as φij,
When Xij ≦ L1, φij = 00 (binary number) (Expression 13)
When L1 <Xij ≦ L2, φij = 01 (binary number) (Expression 14)
When L2 <Xij φij = 10 (binary number) (Equation 15)

In step ST7, if the selected pixel is a pixel to be subjected to transmission processing, a code indicating that the pixel is to be subjected to transmission processing is assigned regardless of the gradation level of the selected pixel (step ST9). In the case of the code assignment example given in step ST8, the code assigned in step ST9 is “11” (binary number) that is not used in the example of step ST8.
φij = 11 (binary number) (Expression 16)

Next, it is confirmed whether or not there is an unprocessed pixel in the selected block (step ST10). If there are unprocessed pixels, the operation from step ST6 is repeated. If there is no unprocessed pixel, the process proceeds to step ST11.
In step ST11, it is confirmed whether there is an unprocessed block in the input image data. If there is an unprocessed block, the operation from step ST2 is repeated. If there is no unprocessed pixel, the process proceeds to step ST12.

  In a state where the processing for all the blocks and all the pixels is completed, the encoded data generating means 6 combines the pixel reference level LA and the pixel value fluctuation width LD for each block and the code φij for each pixel to generate encoded data. (Step ST12). FIG. 5 shows a configuration of encoded data for each block in the present embodiment.

As described above, in this embodiment, when a code is assigned to each pixel divided into blocks, the individual code indicating the pixel of the transmission process is used for the pixel for which the transmission process is performed. It is possible to create encoded data corresponding to the processing.
In addition, since one of the types of codes used when the transparency processing is not considered is used as a code indicating the pixel corresponding to the transparency processing, the data length of the encoded data is changed even when the transparency processing is considered. It is possible to keep the same constant value as in the case of not considering. When the data width of the input image is 8 bits, the same 8 bits as the data width of the input data are assigned to the pixel reference level LA and the pixel value fluctuation width LD of each block, and the encoded data for each pixel is 2 bits. Assign. The data amount per block is 8 bits × 16 pixels = 16 bytes before compression, and after compression is 8 bits + 8 bits + 2 bits × 16 pixels = 6 bytes, so the compression rate is 3/8.

Embodiment 2. FIG.
An embodiment will be described in which 1 bit is assigned instead of 1 gradation in a quantized code as a flag indicating a pixel to be subjected to transparency processing. The configuration diagram showing the image coding apparatus in this case is the same as that in FIG. 1, and the contents of the processing performed by the quantization threshold value calculation means 4 and the quantization means 5 are different from those of the first embodiment.
FIG. 6 is a flowchart showing an image processing method in the image coding apparatus according to the second embodiment of the present invention.

Hereinafter, the processing method of the present embodiment will be described mainly with respect to processing different from that of the first embodiment.
The quantization threshold value calculation means 4 calculates a threshold value L1 for converting the gradation level of each pixel into a code (step ST5). The number of threshold values needs to correspond to the number of bits assigned to the code for converting the gradation level. In the present embodiment, since one bit of the code assigned to each pixel is used as a flag for determining whether or not to perform the transparency process, the code assigned to gradation level conversion for one pixel is set to 1 bit. A case will be described in which the code amount to be assigned is 2 bits as in the first embodiment.

When the code assigned to a pixel is 1 bit, two types of states can be assigned to each pixel. One threshold L1 is required as a threshold for dividing the gradation level into two ranges corresponding to two types of codes. When the threshold value is one, the value of the pixel reference level LA that is the center of the pixel value distribution is used as L1.
L1 = LA (Expression 17)
Here, the code amount assigned to each pixel is 2 bits and the code assigned to gradation level conversion is 1 bit. However, the code amount assigned to each pixel is 3 bits, and the code assigned to gradation level conversion is It is also possible to adopt an encoding method of 2 bits. In this case, since four types of states can be assigned to each pixel, it is necessary to calculate three values as threshold values.

In the quantization means 5, it is confirmed whether or not the pixel selected in step ST6 is a pixel to be subjected to transmission processing (step ST7). It is assumed that pixel data to be subjected to transparency processing in the input image is determined in advance together with the input image.
If the selected pixel is not a pixel to be subjected to the transparent process in step ST7, a 1-bit transparent process corresponding flag corresponding to the selected pixel is set to 0 (step ST13a).
In step ST7, if the selected pixel is a pixel to be subjected to the transparent process, a 1-bit transparent process corresponding flag corresponding to the selected pixel is set to 1 (step ST13b).

Regardless of whether or not the selected pixel is a pixel to be transparently processed, it is determined which region is included in the region constituted by the threshold values obtained in step ST5 corresponding to the gradation level of the selected pixel, and a code is assigned (Step ST8). An example of an allocation method when the number of bits allocated to the code is 1 bit is shown below. The gradation level Xij of the selected pixel and the corresponding code as φij,
When Xij ≦ L1, φij = 0 (binary number) (Equation 18)
When L1 <Xij φij = 1 (binary number) (Equation 19)
The operation of other parts is the same as that in FIG. 2 in the first embodiment.
FIG. 7 shows an encoding threshold distribution and a signal level distribution at the time of decoding when Expression 17 is applied. FIG. 8 shows the configuration of the encoded data for each block in the present embodiment.

  As described above, in the present embodiment, when assigning a code to each pixel divided into blocks, the transparency processing support flag and the code assigned to the conversion of the gradation level are independently held. The pixel value can also be referred to for the corresponding pixel. Therefore, for example, it is possible to cope with a process in which the transmissivity of the transmissive pixel is not set to 100% and the translucent overlay process is performed on the base image with the transmittance of 50%.

Embodiment 3 FIG.
An embodiment in the case where it is determined whether or not each block includes a pixel that transmits data and the coding processing method is switched will be described. The block diagram showing the image encoding apparatus in this case is the same as that shown in FIG. 1, and the contents of the processing performed by the encoding preparation means 3, the quantization threshold calculation means 4, the quantization means 5 and the encoded data creation means 6 are implemented. This is different from the first form.

FIG. 9 is a flowchart showing an image processing method in the image coding apparatus according to Embodiment 3 of the present invention. The operation of the image coding apparatus in FIG. 9 will be mainly described in processing different from that in the first embodiment.
The encoding preparation means 3 calculates the maximum value Lmax and the minimum value Lmin of the pixel gradation level in the block (step ST3), and the pixel reference level LA and the pixel value variation of the block from the distribution of the pixel gradation level in the block. The width LD is obtained (step ST4). The calculation method of the pixel reference level LA and the pixel value variation width LD is the same as that in the first embodiment.
At this time, the encoding preparation means 3 checks whether or not the pixel to be subjected to the transmission process is included in the corresponding block (step ST14).
When a pixel to be subjected to transmission processing is included in the block in step ST14, the quantization threshold value calculation means 4 assigns three types of codes to the gradation levels as in the first embodiment, so that the threshold value L1 , L2 are calculated (step ST5). The calculation method is the same as (Equation 5) and (Equation 6) in the first embodiment.

In step ST14, when the pixel to be subjected to the transparency process is not included in the block, the quantization threshold calculation unit 4 assigns four types of codes to the gradation level without considering the codes for the transparency process. . Therefore, three values L1, L2, and L3 are calculated as threshold values (step ST15). An example of an arithmetic expression is shown below from (Expression 20) to (Expression 22).
L1 = LA-LD / 4 (Formula 20)
L2 = LA (Formula 21)
L3 = LA + LD / 4 (Formula 22)

  In addition, when a pixel to be subjected to the transmission process is included in the block in step ST14, the process in the quantization unit 5 is the same as that in the first embodiment. Steps ST6 to ST10 in the flowchart in FIG. 9 correspond to steps ST6 to ST10 in the flowchart in FIG.

  When the pixel to be subjected to the transparency process is not included in the block in step ST14, the process in the quantization unit 5 is the same as the conventional encoding method that does not consider the transparency process. In step ST17, it is determined in which region the gradation level of the corresponding pixel is included among the four types of regions composed of the threshold values L1, L2, and L3 obtained in step ST15, and a code is assigned. An example of the assignment method when the number of bits assigned to the code is 2 bits is shown in (Expression 23) to (Expression 26) below. The gradation level Xij of the selected pixel and the corresponding code as φij,

When Xij ≦ L1, φij = 00 (binary number) (Equation 23)
When L1 <Xij ≦ L2, φij = 01 (binary number) (Equation 24)
When L2 <Xij ≦ L3 φij = 10 (binary number) (Equation 25)
When L3 <Xij φij = 11 (binary number) (Equation 26)

  The encoded data creation means 6 creates the encoded data by combining the pixel reference level LA and the pixel value variation width LD for each block and the code φij for each pixel in a state where the processing for all the blocks and all the pixels is completed. (Step ST12).

  At this time, the value LD indicating the pixel fluctuation range is stored as a 7-bit value obtained by deleting the least significant 1 bit from the 8-bit value, and a flag indicating whether or not the remaining block corresponds to the transparent processing in the corresponding block. Assign. FIG. 10 shows the structure of the encoded data for each block in the present embodiment.

  As described above, according to the present embodiment, it is determined whether or not the transmission processing is supported for each block, and the transmission processing is performed in order to assign a code corresponding to the pixel that performs the transmission processing only to the block that needs the transmission processing. For blocks that do not exist, it is possible to cope with the transparent processing only at necessary portions while preventing deterioration in image quality due to a decrease in the number of types of assigned codes.

Embodiment 4 FIG.
FIG. 11 is a block diagram showing an image decoding apparatus according to Embodiment 4 of the present invention. In FIG. 11, reference numeral 11 denotes encoded data dividing means for dividing the encoded data into blocks corresponding to image data divided into M × N blocks (M and N are natural numbers), and 12 is included in the encoded data. Gradation level calculation means for calculating a gradation level corresponding to the quantized code from the pixel reference level LA and the pixel value fluctuation width LD for each block, and 13 is quantized for each pixel in the block The gradation level assigning means 14 assigns the gradation levels corresponding to the codes, and image data creating means 14 collects the assigned gradation levels of the respective pixels as decoded image data.

Next, the operation will be described. FIG. 12 is a flowchart showing an image processing method in the image decoding apparatus according to the fourth embodiment of the present invention.
First, the encoded data dividing means 11 divides the encoded data into corresponding portions of each block obtained by dividing the original image data by M × N (M and N are natural numbers) (step ST41). Here, a case where each block is composed of 16 pixels of 4 horizontal pixels and 4 vertical pixels will be described.
One block that has not been decoded is selected from the divided block-unit encoded data (step ST42).

The gradation level calculation means 12 refers to the pixel reference level LA and the pixel value variation width LD for each block from the selected block-unit encoded data, and based on the values of the pixel reference level LA and the pixel value variation width LD. The gradation levels Q1, Q2, and Q3 of the pixel corresponding to the quantized code are calculated (step ST43). An example of a specific calculation method is shown in (Expression 27) to (Expression 29) below.
Q1 = LA-LD / 2 (Formula 27)
Q2 = LA (Formula 28)
Q3 = LA + LD / 2 (Formula 29)

The calculation method corresponds to the case where the quantization threshold is calculated by (Equation 5) and (Equation 6) of the first embodiment. The encoding threshold distribution and the signal level distribution at the time of decoding in this case are as shown in FIG.
One of the codes corresponding to the unprocessed pixels in the selected block is selected (step ST44).

The gradation level assigning means 13 confirms whether or not the selected pixel is a pixel to be subjected to transmission processing (step ST45). Whether or not to perform the transmission process is selected depending on whether or not the reference code φij to be referred to is a specific code indicating a pixel to be subjected to the transmission process. Here, in accordance with the description of (Equation 13) to (Equation 16) in the first embodiment, when the code is 00, 01, 10 (binary number), the transparent process is not performed, and the code is 11 (binary number). Shall be transparent.
If the selected pixel is not a pixel to be subjected to the transmission process in step ST45, the gradation level corresponding to the code is assigned as the gradation level of the selected pixel (step ST46). An example of an assignment method when assigning gradation levels corresponding to (Equation 13) to (Equation 15) of the first embodiment, where the gradation level of the selected pixel is Xij, is shown below (Equation 30) to (Equation 32). Shown in

When φij = 00 (binary number) Xij = Q1 (Equation 30)
When φij = 01 (binary number) Xij = Q2 (Equation 31)
When φij = 10 (binary number) Xij = Q3 (Expression 32)

If the selected pixel is a pixel to be transmitted in step ST45, the gradation level at the corresponding coordinate position of the background image is assigned as the gradation level of the selected pixel (step ST47). Assuming that the gradation level of the background image is Yij, the expression indicating the gradation assignment method is as follows (Expression 33).
When φij = 11 (binary number) Xij = Yij (Expression 33)

It is confirmed whether or not there is a code corresponding to an unprocessed pixel in the selected block (step ST48). If there are unprocessed pixels, the operation from step ST44 is repeated. If there is no unprocessed pixel, the process proceeds to step ST49.
It is confirmed whether there is any unprocessed block-unit encoded data in the input image data (step ST49). If there is an unprocessed block, the operation from step ST42 is repeated. If there is no unprocessed pixel, the process proceeds to step ST50.
The image data creation unit 14 collects the gradation levels of the assigned pixels in a state where the decoding process for all the blocks and all the pixels is completed, and generates decoded image data (step ST50).

  As described above, in the present embodiment, when performing decoding, decoding is performed by performing fine transparency processing in units of pixels in order to refer to background image data corresponding to individual codes indicating pixels to be subjected to transparency processing. It is possible to create converted image data.

Embodiment 5 FIG.
A description will be given of an embodiment of an image decoding processing apparatus in the case of corresponding to encoded data to which one bit is assigned instead of one gradation in a quantized code as a flag indicating a pixel to be subjected to transparent processing. The block diagram showing the image decoding apparatus in this case is the same as that in FIG. 11, and the contents of the processing performed by the gradation level calculation means 12 and the gradation level assignment means 13 are different from those in the fourth embodiment.

FIG. 13 is a flowchart showing an image processing method in the image decoding apparatus according to the fifth embodiment of the present invention. Hereinafter, the processing method of the present embodiment will be described mainly with respect to processing different from that of the fourth embodiment.
The gradation level calculation means 12 refers to the pixel reference level LA and the pixel value variation width LD for each block from the selected block-unit encoded data, and determines the quantum based on the values of the pixel reference level LA and the pixel value variation width LD. The gradation levels Q1 and Q2 of the pixel corresponding to the converted code are calculated (step ST43). In this embodiment, in order to correspond to (Equation 17) of Embodiment 2, a case where a code indicating two types of states is assigned to each pixel will be described. An example of a specific calculation method is shown in (Expression 34) and (Expression 35) below.

Q1 = LA-LD / 2 (Formula 34)
Q2 = LA + LD / 2 (Formula 35)

  The gradation level assigning means 13 confirms whether or not the selected pixel is a pixel to be subjected to transmission processing (step ST45). The reference code φij is composed of 2-bit data, and the lower 1 bit among them is a flag indicating a pixel to be subjected to the transparent process. That is, when the code is 00 or 10 (binary number), the transparent process is not performed, and when the code is 01 or 11 (binary number), the transparent process is performed.

  If the selected pixel is not a pixel to be subjected to the transmission process in step ST45, the gradation level corresponding to the code is assigned as the gradation level of the selected pixel (step ST46). Here, the code corresponding to the gradation level assignment is the upper 1 bit of φij. An example of an allocation method when the gradation level of the selected pixel is Xij and gradation allocation is performed in accordance with (Equation 18) and (Equation 19) of the second embodiment is shown below (Equation 36) and (Equation 37). Shown in

When φij = 00 (binary number) Xij = Q1 (Expression 36)
When φij = 10 (binary number) Xij = Q2 (Expression 37)

If the selected pixel is a pixel to be subjected to transmission processing in step ST45, the gradation level of the selected pixel is obtained by weighted averaging the gradation level assigned to the selected pixel and the gradation level at the corresponding coordinate position of the background image. Is calculated (step ST47).
First, the gradation levels Q1 and Q2 assigned to the selected pixel are calculated by the following (Equation 38) and (Equation 39) as in Step ST46.
When φij = 01 (binary number) X′ij = Q1 (Equation 38)
When φij = 11 (binary number) X′ij = Q2 (Equation 39)
Further, the gradation level at the corresponding coordinate position of the background image is referred to and recorded as a gradation level value.
Assuming that the gradation level of the background image is Yij, the expression indicating the gradation assignment method is as follows (Expression 40).
X ″ ij = Yij (Formula 40)

The two tone level values are weighted and averaged to determine the final tone level of the selected pixel. The calculation formula is as follows (Formula 41).
Xij = αX′ij + (1−α) X ″ ij (Formula 41)
Here, α is a coefficient satisfying 0 ≦ α ≦ 1, and it is assumed that the value is determined in advance from the outside in accordance with the transmittance to be applied to the portion to be subjected to the transmission processing.

  As described above, in the present embodiment, when performing decoding, the gradation level based on the code assigned to the pixel and the gradation level based on the reference of the background image are obtained for the pixel to be subjected to the transparency process, and these are calculated. The final pixel gradation level is calculated by weighted averaging. For this reason, it is possible to realize not only a complete transmission process in which only the background image data is valid, but also a translucent process in which the image subjected to the transmission process can be checked lightly for the pixels for which the transmission process is designated. Is possible.

Embodiment 6 FIG.
An embodiment in the case where it is determined whether or not each block includes a pixel that transmits data and the decoding processing method is switched will be described. The block diagram showing the image decoding apparatus in this case is the same as that in FIG. 11, and the contents of the processing performed by the gradation level calculation means 12 and the gradation level assignment means 13 are different from those in the fourth embodiment.
FIG. 14 is a flowchart showing an image processing method in the image decoding apparatus according to Embodiment 6 of the present invention. Hereinafter, the processing method of the present embodiment will be described mainly with respect to processing different from that of the fourth embodiment.

The gradation level calculation means 12 first refers to a 1-bit transmission-compatible block flag from the selected block-unit encoded data, and checks whether or not a pixel to be subjected to transmission processing is included in the block (step ST51).
If the block selected in step ST51 includes a pixel to be transparently processed, gradation levels Q1, Q2, and Q3 are assigned to the three types of codes as in the fourth embodiment. The calculation method is the same as (Equation 27) to (Equation 29) in the fourth embodiment (step ST43). However, since the present embodiment corresponds to the decoding of the encoded data in the third embodiment, the pixel value variation width LD value is transmitted when the pixel to be subjected to the transmission processing is not included in the block in step ST51. Since there is no need to consider codes for processing, there are four types of codes assigned to gradation levels. For this reason, four values of gradation levels Q1, Q2, Q3, and Q4 corresponding to the four codes are calculated (step ST52).

An example of arithmetic expressions (Expression 42) to (Expression 45) in the case of corresponding to the encoding of (Expression 20) to (Expression 22) in the third embodiment is shown below.
Q1 = LA-LD / 2 (Formula 42)
Q2 = LA-LD / 8 (Formula 43)
Q2 = LA + LD / 8 (Formula 44)
Q3 = LA + LD / 2 (Formula 45)
In addition, when the pixel to be subjected to the transmission process is included in the block in step ST51, the process in the gradation level assigning unit 4 is the same as that in the fourth embodiment. Steps ST44 to ST48 in the flowchart of FIG. 14 correspond to steps ST44 to ST48 in the flowchart of FIG.

  When the pixel to be subjected to the transparent process is not included in the block in step ST51, the process in the gradation level assigning unit 13 is the same as the conventional decoding method that does not consider the transparent process. In step ST54, the gradation level corresponding to the code is assigned as the gradation level of the selected pixel from the gradation levels Q1, Q2, Q3, and Q4 corresponding to the four codes obtained in step ST52. An example of an assignment method when assigning gradations corresponding to (Equation 13) to (Equation 15) of the third embodiment, where the gradation level of the selected pixel is Xij, from (Equation 46) to (Equation 49) below. Shown in

When φij = 00 (binary number) Xij = Q1 (Equation 46)
When φij = 01 (binary number) Xij = Q2 (Formula 47)
When φij = 10 (binary number) Xij = Q3 (Formula 48)
When φij = 11 (binary number) Xij = Q4 (formula 49)
It is confirmed whether there is any unprocessed block-unit encoded data in the input image data (step ST49). If there is an unprocessed block, the operation from step ST42 is repeated. If there is no unprocessed pixel, the process proceeds to step ST50.
The image data creation means 14 generates image data that is decoded together with the gradation levels of the assigned pixels in a state where the decoding process for all blocks and all pixels is completed (step ST50).

  As described above, in the present embodiment, it is determined whether or not to support transparency processing based on a flag assigned to each block, and decoding corresponding to pixels that perform transparency processing only for blocks that require transparency processing. Since the processing is performed, it is possible to cope with the transparent processing only in necessary portions while preventing the image quality from being deteriorated due to the decrease in the number of assigned code types for the blocks that are not subjected to the transparent processing.

Embodiment 7 FIG.
In the first to sixth embodiments, it is assumed that the base image for performing the transparency process is prepared in advance. In the seventh embodiment, a case will be described in which the background image is encoded and decoded in the same manner as the overlay image for which the transparent process is performed. In this case, the value LA indicating the pixel reference level at the time of encoding is stored as a 7-bit value obtained by deleting the least significant 1 bit from the 8-bit value, and whether or not the corresponding image is a background image is stored in the remaining 1 bit. Assign a flag.

The encoding process includes block division means, signal level range calculation means, encoding preparation means, quantization threshold value calculation means, quantization means, and encoded data in the block diagram shown in FIG. Using the creation means, the encoded data is created by the same processing as in the case of the pixels that are not subjected to the transparency processing in the first embodiment.
In addition, as described above, 1 bit of the pixel value reference level signal is used as a flag for determining whether the data for the superimposition image or the background image is the bit data of the pixel value reference level signal.
FIG. 15 shows a configuration of encoded data for each block in the present embodiment.

The configuration of the image decoding apparatus is the same as that shown in FIG. 11, and in the decoding process, the background image data is divided and encoded for each M × N (M and N are natural numbers) blocks, and the pixels in the block are encoded. Encoded data in which the reference level, the pixel value fluctuation range, and the code of each pixel are collected is input in addition to the encoded data of the overlay image, and the encoded data of the background image and the above-described overlay In the encoded image data, one bit of each pixel value reference level signal is used as a flag for discriminating whether the data is an overlay image data or a background image data.
The input encoded data refers to the above-mentioned flag (the flag indicating whether the corresponding image is the background image using the least significant 1 bit of the pixel reference level LA), and whether the encoded data corresponds to the background image is subjected to a transparent process. It is determined whether it corresponds to the image to be performed. If the input encoded data is a background image, the data is temporarily stored in the buffer area provided, and the background image data stored in the buffer area when the overlay image for transparency processing is input And the decoded image data is generated.

  As described above, according to the present embodiment, it is possible to determine whether the image data is a background image or a superimposition image to be subjected to transparency processing based on the encoded data flag. Any image data can be processed using the same data input method without preparing a method for reading and holding data.

  The present invention is a technique related to encoding / decoding of image processing for superimposing images, and can be mounted on image processing products such as in-vehicle devices and surveillance camera devices.

It is a block diagram which shows the image coding apparatus which is Embodiment 1 of this invention. It is a flowchart which shows the image processing method in the image coding apparatus by Embodiment 1 of this invention. It is a figure which shows the signal level distribution in the conventional encoding system which matches the gradation level of each pixel with four types of codes. It is a figure which shows the encoding threshold value distribution at the time of dividing | segmenting a gradation level into three ranges, and the signal level distribution at the time of decoding. FIG. 3 is an encoded data configuration diagram for each block in the first embodiment. It is a flowchart which shows the image processing method in the image coding apparatus by Embodiment 2 of this invention. It is a figure which shows the encoding threshold value distribution at the time of dividing | segmenting the gradation level by Embodiment 2 into two ranges, and the signal level distribution at the time of decoding. FIG. 10 is a block diagram of encoded data for each block in the second embodiment. 10 is a flowchart illustrating an image processing method in the image encoding device according to the third embodiment. It is a figure which shows the structure of the encoding data for every block in Embodiment 3. FIG. [Fig. 10] Fig. 10 is a configuration diagram illustrating an image decoding device according to a fourth embodiment. 14 is a flowchart illustrating an image processing method in the image decoding apparatus according to the fourth embodiment. 16 is a flowchart illustrating an image processing method in the image decoding apparatus according to the fifth embodiment. 20 is a flowchart illustrating an image processing method in the image decoding apparatus according to the sixth embodiment. FIG. 25 is a diagram showing a configuration of encoded data for each block in the seventh embodiment.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1; Block division means, 2; Signal level range calculation means, 3; Encoding preparation means, 4; Quantization threshold value calculation means, 5; Quantization means, 6: Encoded data preparation means, 11; Encoded data division Means 12; gradation level calculation means 13; gradation level assignment means 14; image data creation means.

Claims (4)

Block dividing means for dividing the input image data of the overlay image into M × N (M and N are natural numbers) blocks, and signal level range calculating means for calculating the maximum and minimum pixel values in the block Encoding preparation means for calculating the reference level of the pixel of the block and the pixel value fluctuation range in the block from the maximum value and the minimum value of the obtained pixel value, and the maximum value and the minimum value of the pixel value obtained in the same manner Quantization threshold value calculation means for calculating a quantization threshold value of a pixel value for dividing the gradation level corresponding to the number of types of codes to be converted, and the gradation level of each pixel in the block as the quantization threshold value Depending on which of the plurality of ranges determined in accordance with each range, the gradation level of each pixel is replaced with a unique code corresponding to each range, and when a plurality of images are overlaid, transmission is performed. For the pixels that output the background image data, the quantization means for assigning a unique code of a different type from the above codes, the reference level of the pixels in the block, the pixel value fluctuation range, and the codes assigned to the individual pixels are summarized. An encoded data creating means for producing encoded data ,
The encoding preparation means calculates and outputs the reference level of the pixel of the block and the pixel value fluctuation range in the block, and outputs a signal indicating whether or not the block includes a pixel that transmits data in the block. Output included in the fluctuation range signal,
The quantization threshold value calculation means, when a pixel that transmits data is included in the block, a threshold value for dividing the gradation level into n stages, and when a pixel that transmits data is not included in the block, Each of the threshold values for dividing the gradation level into (n + 1) stages is obtained,
When the block includes pixels that transmit data, the quantization unit assigns a code indicating either a transparent pixel or a code indicating n gradation levels for each pixel, and assigns the data to the block. An image coding apparatus characterized by assigning a code indicating a gradation level of (n + 1) stages for each pixel when a transparent pixel is not included. Data M × N of the input background image (M, N is a natural number) blocks dividing means for dividing a block of a signal level range calculating means for calculating the maximum value and the minimum value of pixel values in the block, obtained Encoding preparation means for calculating the reference level of the pixel of the block and the fluctuation range of the pixel value in the block from the maximum value and the minimum value of the obtained pixel value, and a code for converting from the maximum value and the minimum value of the obtained pixel value Quantization threshold value calculation means for calculating a quantization threshold value of a pixel value for dividing the gradation level corresponding to the number of types, and the gradation level of each pixel in the block is determined by the quantization threshold value. by either included in any of the range of the plurality of ranges that, quantizing means for replacing the gray level of each pixel in the specific codes corresponding to the respective ranges, the reference level of the interblock-pixel pixel value variance Width, an apparatus and a coded data generating means to collectively code assigned to each pixel coded data,
Use Irare the flag indicating that the data of the lower ground image one bit of the pixel value reference level signal, a base image used in the encoded superimposed image by the image encoding apparatus according to claim 1, wherein picture picture encoding system. The image data of the image for superimposition is divided and encoded for each M × N (M and N are natural numbers) blocks, and the reference level of the pixels in the block, the pixel value fluctuation range, and a plurality of images are superimposed and combined. An image decoding apparatus that inputs encoded data in which different types of codes assigned to pixels when a plurality of images are not overlapped with a pixel that is transmitted through and output the background image data and decodes the image data. There,
Gradation level calculation means for calculating the gradation level corresponding to the code of each pixel from the pixel reference level and pixel value fluctuation range of each block, and gradation level assignment means for replacing the code of each pixel with the gradation level Have
The gradation level calculation means calculates the gradation level corresponding to the code of each pixel from the pixel reference level and pixel value variation width of each block, and is included in the pixel value variation value width data in the block encoded data. It is determined whether the corresponding block includes a pixel that transmits data according to the flag, and if the corresponding block includes a pixel that transmits data, the gradation level corresponding to n types of codes is calculated and the corresponding block is calculated. Does not include pixels that transmit data, the gradation level corresponding to (n + 1) types of codes is calculated,
In the case where the block includes pixels that transmit data, if the code of the pixel indicates a transmission output, the gradation level assigning unit sets the gradation level of the pixel to the background of another image. In addition to replacing with the corresponding pixel gradation level , in other cases, the code of each pixel is replaced with the gradation level according to the correspondence between the n types of codes calculated by the gradation level calculating means and the gradation level, An image characterized by replacing the code of each pixel with a gradation level according to the correspondence between the (n + 1) types of codes calculated by the gradation level calculation means and the gradation level when the pixel that does not transmit data is included Decryption device. In addition to the encoded data of the overlay image, the background image data is divided and encoded for each block of M × N (M and N are natural numbers), and the reference level of the pixels in the block, the pixel value variation width, coded data code pixels are together is further inputted, the encoded data of the background image, 1-bit pixel value reference level signal is used to flag to determine the data of the lower fabric image,
If the input encoded data is a background image, the data is temporarily stored in the buffer area provided with the data, and the background image data stored in the buffer area when the overlay image for transparent processing is input The image decoding apparatus according to claim 3 , wherein the decoding is performed by a combination with.

Images