JP2009071472A - Image encoding method, image decoding method, image encoder, image decoder and semiconductor integrated circuit - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image encoder or the like capable of accurately encoding images even when a strong edge part is present. <P>SOLUTION: The image encoder 10 includes: a first difference absolute value calculation part 21 for calculating the absolute value dH of a difference between a pixel C' positioned at the upper left of an encoding object pixel D and a pixel B' positioned above the pixel D; a second difference absolute value calculation part 22 for calculating the absolute value dV of a difference between the pixel C' and a pixel A' positioned at the left of the pixel D; a difference calculation part 23 for calculating a difference dF between the dH and the dV; a predicted value calculation part 24 for calculating the predicted value P of the pixel D on the basis of the dF; a pixel predicted value difference calculation part 25 for calculating a difference d between the pixel D and the predicted value P; and a quantization part 26 for quantizing the difference d. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to an image encoding method for encoding image data and an image decoding method for decoding image data encoded by such an image encoding method. The present invention also relates to an image encoding apparatus that encodes image data and an image decoding apparatus that decodes image data encoded by such an image encoding apparatus. Furthermore, the present invention relates to a semiconductor integrated circuit including such an image encoding device and an image decoding device.

  Conventionally, an encoding target pixel has been encoded based on pixels around the encoding target pixel. Such image encoding will be described with reference to FIG.

As shown in FIG. 9, when the pixel located on the left of the encoding target pixel X is A, the pixel located on the encoding target pixel X is B, and the pixel located on the upper left of the encoding target pixel X is C. , The predicted value X ′ of the encoding target pixel X is
X ′ = A / 2 + (C + B) / 4 (1)
It is calculated by.

  However, when a strong edge portion such as an end portion of text exists between the pixels A, B, C, and X, the predicted value X ′ calculated by the above equation (1) is the encoding target pixel. The value may be greatly deviated from X.

  Further, in Patent Document 1 below, in an image compression apparatus that performs lossless image compression in which original information is stored through a compression / expansion process, a digital image signal is input and is located around a prediction target pixel. Pixel correlation evaluation means for evaluating individual correlations of referenceable pixels, optimum value calculation means for calculating an optimum adaptive processing coefficient that maximizes correlation evaluation between pixels, optimum adaptive processing coefficient, and prediction target pixel And an entropy encoding means for entropy encoding the DPCM encoded image signal. The image compression apparatus includes a DPCM encoding means for generating a DPCM encoded image signal from the above and an entropy encoding means for entropy encoding the DPCM encoded image signal.

  However, in order to evaluate the individual correlation of the referenceable pixels around the prediction target pixel and to calculate the optimum adaptive processing coefficient that maximizes the correlation evaluation between the pixels, the process becomes complicated. The amount of processing increases, and the memory capacity required to store the reference pixels increases. Furthermore, in some cases, a code for identification is required, and the amount of codes increases.

JP 2001-77700 A

  The present invention has been made in view of the technical problems as described above, and an object of the present invention is to ensure that an image is accurately processed with simple processing and configuration even when a strong edge portion is present. It is an object of the present invention to provide an image encoding method and an image encoding apparatus that can be encoded. The present invention also provides an image decoding method and an image decoding apparatus for decoding an image encoded by such an image encoding method and an image encoding apparatus. Another object of the present invention is to provide a semiconductor integrated circuit including such an image encoding device and an image decoding device.

  The present invention encodes an encoding target pixel in image data including a plurality of pixels in a first direction adjacent to the encoding target pixel in the first direction and in a second direction intersecting the first direction. An image code to be encoded based on a second pixel adjacent to the pixel to be converted and a third pixel adjacent to the first pixel in the second direction and adjacent to the second pixel in the first direction When the absolute value of the difference between the third pixel and the second pixel is larger than the absolute value of the difference between the third pixel and the first pixel, the second pixel is changed to the first pixel. If the absolute value of the difference between the third pixel and the second pixel is smaller than the absolute value of the difference between the third pixel and the first pixel with a weight greater than the second pixel, The encoding target image is obtained by performing a weighting operation on the first pixel and the second pixel with a weight smaller than one pixel. And of calculating the predicted value, to encode the encoding target pixel by quantizing the difference between the coded pixel and the predicted value, related to the image coding apparatus.

  According to the present invention, it is possible to accurately encode an image even when a strong edge portion exists.

  According to the present invention, an encoding target pixel in image data including a plurality of pixels is divided into a first pixel adjacent to the encoding target pixel in the first direction and a second direction intersecting the first direction. Are encoded based on the second pixel adjacent to the encoding target pixel and the third pixel adjacent to the first pixel in the second direction and adjacent to the second pixel in the first direction. In the image encoding method, the step (a) of calculating the absolute value of the difference between the third pixel and the second pixel, and the absolute value of the difference between the third pixel and the first pixel are calculated. Step (b) and a step of calculating a weighting coefficient representing weighting of the first pixel and the second pixel with respect to the encoding target pixel, wherein an absolute value of a difference between the third pixel and the second pixel is If the absolute value of the difference between the third pixel and the first pixel is greater than the second pixel, A weighting coefficient that calculates a weight greater than the first pixel is calculated, and the absolute value of the difference between the third pixel and the second pixel is smaller than the absolute value of the difference between the third pixel and the first pixel. In this case, the step (c) of calculating a weighting coefficient that sets the second pixel to be smaller in weight than the first pixel, and performing the weighting operation on the first pixel and the second pixel using the weighting coefficient. To calculate the prediction value of the encoding target pixel (d), to calculate the difference between the encoding target pixel and the prediction value (e), and to quantize the value calculated in step (e) Step (f).

  According to the present invention, it is possible to accurately encode an image even when a strong edge portion exists.

  Further, in the present invention, the step (c) includes the step of calculating the difference between the absolute value of the difference between the third pixel and the second pixel and the absolute value of the difference between the third pixel and the first pixel. If greater than the first threshold, the given first value is a weighting factor, and the absolute value of the difference between the third pixel and the second pixel and the difference between the third pixel and the first pixel If the difference from the absolute value of the second pixel is smaller than the given second threshold value, the given second value is used as a weighting factor, and the absolute value of the difference between the third pixel and the second pixel is equal to the third value. When the difference between the absolute value of the difference between the first pixel and the first pixel is smaller than the first threshold and larger than the second threshold, the first value and the second value are interpolated. The obtained value may be used as a weighting coefficient.

  In this way, the weighting coefficient can be calculated by a simple calculation.

  In the present invention, each of the plurality of pixels may include a plurality of components, and steps (a) to (f) may be performed for each of the components of the encoding target pixel.

  In this way, even if a pixel includes a plurality of components, an image can be encoded accurately.

  In the present invention, each of the plurality of pixels includes an R component, a G component, and a B component. For each of the R component, the G component, and the B component of the encoding target pixel, steps (a) to (step) are performed. (F) may be performed.

  In the present invention, the step (g) for dequantizing the value calculated in step (f), the predicted value calculated in step (d), and the value calculated in step (g) And a step (h) of storing the value added to the buffer in the buffer, and the value stored in the buffer in the step (h) is set to the fourth direction adjacent to the encoding target pixel in the direction opposite to the first direction. The fifth pixel adjacent to the pixel to be encoded in the direction opposite to the second direction and / or the fourth pixel in the direction opposite to the second direction and the direction opposite to the first direction. May be used when encoding a sixth pixel adjacent to the fifth pixel.

  In addition, the present invention provides a decoding target pixel in image data encoded by the image encoding method according to the present invention, a first pixel adjacent to the decoding target pixel in the first direction, A second pixel adjacent to the pixel to be decoded in a second direction intersecting the direction, and a third pixel adjacent to the first pixel in the second direction and adjacent to the second pixel in the first direction. An image decoding method for decoding based on a pixel, the step (a) of calculating an absolute value of a difference between a third pixel and the second pixel, and the third pixel and the first pixel Calculating an absolute value of a difference between the first pixel and the second pixel, and calculating a weighting coefficient representing a weighting of the first pixel and the second pixel with respect to the encoding target pixel, the third pixel and the second pixel The absolute value of the difference from the pixel is the difference between the third pixel and the first pixel Is greater than the absolute value of the first pixel, a weighting coefficient is calculated that sets the second pixel as a weight greater than the first pixel, and the absolute value of the difference between the third pixel and the second pixel is When the absolute value of the difference from the first pixel is smaller than the first pixel, the step (c) of calculating a weighting coefficient that sets the second pixel as a weight smaller than the first pixel, Performing a weighting operation on the pixel and the second pixel to calculate a prediction value of the decoding target pixel, a step (e) of dequantizing the decoding target pixel, and a step (d) And the step (f) of adding the predicted value calculated in step (e) to the value calculated in step (e).

  According to the present invention, image data encoded by the image encoding method according to the present invention can be accurately decoded.

  According to the present invention, an encoding target pixel in image data including a plurality of pixels is divided into a first pixel adjacent to the encoding target pixel in the first direction and a second direction intersecting the first direction. Are encoded based on the second pixel adjacent to the encoding target pixel and the third pixel adjacent to the first pixel in the second direction and adjacent to the second pixel in the first direction. An image encoding device, a first difference absolute value calculation unit that calculates an absolute value of a difference between a third pixel and a second pixel, and an absolute value of a difference between the third pixel and the first pixel The difference between the absolute value of the difference calculated by the second difference absolute value calculation unit, the difference absolute value calculated by the first difference absolute value calculation unit, and the absolute value of the difference calculated by the second difference absolute value calculation unit is calculated. And a prediction value of the pixel to be encoded is calculated based on the difference calculated by the difference calculation unit and the difference calculated by the difference calculation unit. A prediction value calculation unit that calculates a difference between a prediction target pixel and a prediction value calculated by the prediction value calculation unit, and a difference calculated by the pixel prediction value difference calculation unit. The present invention relates to an image encoding device including a quantization unit that performs quantization.

  According to the present invention, it is possible to accurately encode an image even when a strong edge portion exists.

  In the present invention, each of the plurality of pixels includes a plurality of components, and the first difference absolute value calculation unit, the second difference absolute value calculation unit, the difference calculation unit, the prediction value calculation unit, the pixel prediction value difference calculation unit, In addition, a plurality of sets of quantization units may be provided, and each set may encode each component of the encoding target pixel in parallel.

  In this way, even when a pixel includes a plurality of components, an image can be encoded at high speed.

  The present invention also provides a decoding target pixel in image data encoded by the image encoding apparatus according to the present invention, a first pixel adjacent to the decoding target pixel in the first direction, A second pixel adjacent to the pixel to be decoded in a second direction intersecting the direction, and a third pixel adjacent to the first pixel in the second direction and adjacent to the second pixel in the first direction. An image decoding apparatus that performs decoding based on a pixel, a first difference absolute value calculation unit that calculates an absolute value of a difference between a third pixel and a second pixel, a third pixel, and a first pixel A second difference absolute value calculation unit that calculates an absolute value of a difference from the pixel, a difference absolute value calculated by the first difference absolute value calculation unit, and a difference calculated by the second difference absolute value calculation unit Based on the difference calculated by the difference calculator and the difference calculated by the difference calculator A prediction value calculation unit that calculates a prediction value of a decoding target pixel, an inverse quantization unit that inversely quantizes the decoding target pixel, and a prediction value and an inverse quantization unit calculated by the prediction value calculation unit. The present invention relates to an image decoding device including a decoding unit that decodes a decoding target pixel by adding the calculated value.

  According to the present invention, the image data encoded by the image encoding device according to the present invention can be accurately decoded.

  The present invention also provides an image encoding device according to the present invention, a memory for storing image data encoded by the image encoding device, and an image according to the present invention for decoding image data stored in the memory. The present invention relates to a semiconductor integrated circuit including a decoding device.

  According to the present invention, the memory capacity required for storing image data can be reduced.

  Hereinafter, embodiments of the present invention will be described in detail. In addition, this embodiment demonstrated below does not unduly limit the content of this invention described in the claim. Further, not all of the configurations described in the present embodiment are essential as a solution means of the present invention. The same constituent elements are denoted by the same reference numerals and description thereof is omitted.

  FIG. 1 shows a configuration example of an image encoding apparatus according to an embodiment of the present invention. The image encoding device 10 includes an image encoding unit 11 to which an image signal (frame data) is input, and a line buffer (FIFO buffer) 12. Note that the image encoding device 10 may be realized as a semiconductor integrated circuit.

  Here, the frame data input to the image encoding device 10 will be described. In this embodiment, as shown in FIG. 2, one frame data has 320 columns in the main scanning line direction (here X direction) and 240 rows in the sub scanning line direction (here Y direction). Contains pixels. Each pixel of the frame data is an image encoding device in the order of coordinates (0, 0), (1, 0), ..., (319, 0), (0, 1), ..., (319, 239). 10 is sequentially input, and the image encoding device 10 sequentially encodes and sequentially outputs each pixel that is sequentially input.

  In FIG. 2, when the pixel D is an encoding target pixel, the total from the pixel C ′ positioned at the upper left of the encoding target pixel D to the pixel A ′ positioned adjacent to the left of the encoding target pixel D. 321 pixels are stored in the line buffer 12 (see FIG. 1). Note that the pixels stored in the line buffer 12 are not the original pixels (pixel A, pixel B, pixel C,...) Input to the image encoding device 10 from the outside, but are encoded by the image encoding device 10. Further decoded pixels (pixel A ′, pixel B ′, pixel C ′,...). Thus, in the present embodiment, “′” is added to the encoded and further decoded pixels.

  In the present embodiment, each pixel is described as having an 8-bit width and can take a value from 0 to 255. When each pixel includes a plurality of components (for example, RGB, YUV, YCbCr, etc.), the image encoding device 10 may sequentially encode each component of the encoding target pixel. For example, when each pixel includes an RGB component, the image encoding device 10 may sequentially encode the R component, the G component, and the B component of each pixel for each component. Alternatively, the image encoding device 10 includes the same number of sets of the image encoding unit 11 and the line buffer 12 as the number of pixel components, and each set encodes each component of the encoding target pixel in parallel. You may do it. For example, when each pixel includes an RGB component, the image encoding device 10 includes three sets of the image encoding unit 11 and the line buffer 12, and each set includes an R component of the encoding target pixel, G The encoding of the component and the B component may be performed in parallel. In this way, even when a pixel includes a plurality of components, encoding can be performed at high speed.

  When each pixel includes an RGB component, as an example, each pixel has an R (red) component of 8 bits wide, a G (green) component of 8 bits wide, and a B (blue) component of 8 bits wide for a total of 24 bits. It is good also as width.

  Referring to FIG. 1 again, the image encoding unit 11 includes a first difference absolute value calculation unit 21, a second difference absolute value calculation unit 22, a difference calculation unit 23, a predicted value calculation unit 24, and a pixel predicted value. The difference calculation unit 25, the quantization unit 26, the inverse quantization unit 27, and the decoding unit 28 are included.

  The first difference absolute value calculation unit 21 calculates the absolute value of the difference between the pixel located at the upper left of the encoding target pixel and the pixel located above the encoding target pixel.

  The second difference absolute value calculation unit 22 calculates the absolute value of the difference between the pixel located on the upper left of the encoding target pixel and the pixel located on the left adjacent to the encoding target pixel.

The difference calculation unit 23 calculates a difference between the value calculated by the first difference absolute value calculation unit 21 and the value calculated by the second difference absolute value calculation unit 22.
The predicted value calculation unit 24 calculates the predicted value of the encoding target pixel based on the value calculated by the difference calculation unit 23.

The pixel prediction value difference calculation unit 25 calculates a difference between the encoding target pixel and the prediction value calculated by the prediction value calculation unit 24.
The quantization unit 26 quantizes the value calculated by the pixel prediction value difference calculation unit 25. The value quantized by the quantization unit 26 is output to the external circuit at the subsequent stage.
The inverse quantization unit 27 performs inverse quantization on the value calculated by the quantization unit 26.

  The decoding unit 28 decodes the encoding target pixel by adding the value inversely quantized by the inverse quantization unit 27 and the prediction value calculated by the prediction value calculation unit 24. The pixels decoded by the decoding unit 28 are stored in the line buffer 12, and the subsequent pixels (the pixel located on the right side of the current encoding target pixel, the pixel positioned on the lower side of the current encoding target pixel) , The pixel located at the lower right of the current pixel to be encoded).

  Next, the operation of the image encoding device 10 will be described.

  FIG. 3 is a flowchart showing processing executed by the image encoding device 10. The image encoding device 10 executes the process shown in FIG. 3 every time a pixel (encoding target pixel) is input from the outside. If each pixel includes a plurality of components, the processing shown in FIG. 3 may be executed for each component of the decoding target pixel.

First, the first difference absolute value calculation unit 21 sets a pixel (here, pixel C ′ (see FIG. 2)) positioned at the upper left of the pixel to be encoded (here, pixel D (see FIG. 2)). ) And a pixel (here, referred to as a pixel B ′ (see FIG. 2)) positioned above the encoding target pixel D, an absolute value dH of the difference is expressed as dH = | C′−B ′ | 2)
(Step S11).

Next, the second difference absolute value calculation unit 22 sets a pixel C ′ positioned at the upper left of the encoding target pixel D and a pixel positioned adjacent to the left of the encoding target pixel D (here, pixel A ′ (FIG. 2)). Refer to the absolute value dV of the difference with d) = V = | C′−A ′ | (3)
(Step S12).

Next, the difference calculation unit 23 sets the difference dF between the value dH calculated by the first difference absolute value calculation unit 21 and the value dV calculated by the second difference absolute value calculation unit 22 as follows: dF = dH−dV (4)
(Step S13).

  Next, the predicted value calculation unit 24 calculates the predicted value P of the encoding target pixel D based on the value dF calculated by the difference calculation unit 23 (steps S14 to S19).

Specifically, the predicted value calculation unit 24 sets the predicted value P to P = (A ′ + 7B ′) / 8 when dF is larger than a given threshold Th (here, 64) (step S14). ... (5)
(Step S15).

Further, when dF is smaller than a given threshold −Th (here, −64) (step S16), the predicted value calculation unit 24 sets the predicted value P to P = (7A ′ + B ′) / 8. ... (6)
(Step S17).

In addition, when dF is larger than the threshold value −Th and smaller than the threshold value Th, the predicted value calculation unit 24 calculates a weighting coefficient m that represents the weighting of the pixel A ′ and the pixel B ′ with respect to the encoding target pixel D (step). S18). For example, as shown in FIG. 4, when dF = Th, m = 3/8, and when dF = −Th, m = −3 / 8, and a straight line between these two points according to the value of dF. You may make it interpolate. Then, the predicted value calculation unit 24 uses the weighting coefficient m calculated in step S18 to change the predicted value P to P = 4/8 (A ′ + B ′) + m (B′−A ′) (7) )
(Step S19).

  In FIG. 4, when dF> 0, m> 0, that is, the weight of the pixel B ′ is larger than the weight of the pixel A ′. On the other hand, when dF <0, m <0, that is, the weight of the pixel B ′ is smaller than the weight of the pixel A ′.

  In steps S14 to S15, when dF is larger than the threshold Th, calculating the predicted value P by the equation (5) can be considered as calculating the weighting coefficient m as 3/8. Further, in steps S16 to S17, when dF is smaller than the threshold value -Th, calculating the predicted value P by equation (6) may be considered as calculating the weighting coefficient m to -3/8. it can.

Next, the pixel prediction value difference calculation unit 25 calculates the difference d between the encoding target pixel D and the prediction value P as follows: d = D−P (8)
(Step S20).

Then, the quantization unit 26 converts the difference d calculated by the pixel prediction value difference calculation unit 25 to iCode = Quant (d) (9)
(Step S21). Here, Quant () represents a quantization operation. As the quantization operation, various known operations can be used. For example, table conversion using a quantization table can also be used. A fixed-length quantization operation can also be used. The value iCode obtained by the quantization is output as an output image signal to the external circuit at the subsequent stage.

Next, the inverse quantization unit 27 converts the value iCode calculated by the quantization unit 26 to d ′ = deQuant (iCode) (10)
Then, inverse quantization is performed (step S22). Here, deQuant () represents an inverse quantization operation. As the inverse quantization operation, various known operations can be used. For example, table conversion using an inverse quantization table can also be used.

Then, the decoding unit 28 encodes and further decodes the encoding target pixel D based on the predicted value P calculated by the predicted value calculation unit 24 and the value d ′ calculated by the inverse quantization unit 27. The pixel D ′ is D ′ = P + d ′ (11)
(Step S23). This pixel D ′ is stored in the line buffer 12, and the subsequent pixel (the pixel located right next to the current encoding target pixel, the pixel positioned next to the current encoding target pixel, the current encoding target, Used to encode a pixel located at the lower right of the pixel).

Next, an embodiment of the present invention will be described with reference to FIGS.
FIGS. 5A to 5C are diagrams illustrating examples of images input to the image encoding device 10.

  In FIG. 5A, it is assumed that the value of the pixel A ′ is “0”, the value of the pixel B ′ is “255”, and the value of the pixel C ′ is “0”. That is, a strong edge portion in the vertical direction (Y direction) exists between the pixels A ′, B ′, and C ′. Such a strong edge portion may appear at the end of the text. In this case, it is considered that the value of the encoding target pixel D is often close to “255”. That is, it is considered that the encoding target pixel D often has a higher correlation with the pixel B ′ than the pixel A ′. Therefore, when encoding the encoding target pixel D, it is considered preferable to make the weight of the pixel B ′ larger than the pixel A ′.

When such an image is input to the image encoding device 10, the absolute value dH of the difference between the pixel C ′ and the pixel B ′ is expressed by the following equation (2):
dH = | C′−B ′ |
= | 0-255 |
= | -255 |
= 255 (12)
(See step S11).

Further, the absolute value dV of the difference between the pixel C ′ and the pixel A ′ is expressed by the following equation (3):
dV = | C′−A ′ |
= | 0-0 |
= 0 (13)
(See step S12).

And the difference dF between the value dH and the value dV is given by the equation (4):
dF = dH−dV
= 255-0
= 255 (14)
(See step S13).

Since the difference dF is larger than the threshold Th (here, 64) (see step S14), the predicted value P of the encoding target pixel D is expressed by the following equation (5):
P = (A ′ + 7B ′) / 8
= (0 + 7 · 255) / 8
= 1785/8
= 223.125 (15)
(See step S15).

  As described above, when the value of the pixel A ′ is “0”, the value of the pixel B ′ is “255”, and the value of the pixel C ′ is “0”, the encoding target pixel It is considered that the value of D is often close to “255”. Therefore, as shown in the equation (15), the weight of the pixel B ′ is made larger than the pixel A ′, and the predicted value P is set to a value close to “225”, so that the encoding target pixel D and the predicted value P are It is considered that the difference can be reduced. This makes it possible to increase the compression rate while maintaining high image quality.

  The reason why the weight of the pixel B ′ is not set to 100% in the expressions (5) and (15) is as follows. When the value of the pixel A ′ is “0”, the value of the pixel B ′ is “255”, and the value of the pixel C ′ is “0”, the value of the encoding target pixel D is “255”. It is thought that there are many cases close to. However, even in such a case, the value of the encoding target pixel D may not be close to “255”. In view of this, the weight of the pixel B ′ is not set to 100%.

  Next, FIG. 5B will be described. In FIG. 5B, the value of the pixel A ′ is “0”, the value of the pixel B ′ is “255”, and the value of the pixel C ′ is “255”. That is, a strong edge portion in the horizontal direction (X direction) exists between the pixels A ′, B ′, and C ′. Such a strong edge portion may appear at the end of the text. In this case, it is considered that the value of the encoding target pixel D is often close to “0”. That is, it is considered that the encoding target pixel D often has a higher correlation with the pixel A ′ than the pixel B ′. Therefore, when encoding the encoding target pixel D, it is considered preferable to make the weight of the pixel A ′ larger than the pixel B ′.

When such an image is input to the image encoding device 10, the absolute value dH of the difference between the pixel C ′ and the pixel B ′ is expressed by the following equation (2):
dH = | C′−B ′ |
= | 255-255 |
= 0 (16)
(See step S11).

Further, the absolute value dV of the difference between the pixel C ′ and the pixel A ′ is expressed by the following equation (3):
dV = | C′−A ′ |
= | 255-0 |
= | 255 |
= 255 (17)
(See step S12).

And the difference dF between the value dH and the value dV is given by the equation (4):
dF = dH−dV
= 0-255
= -255 (18)
(See step S13).

Since the difference dF is smaller than the threshold value -Th (here, -64) (see step S16), the prediction value P of the encoding target pixel D is expressed by the following equation (6):
P = (7A ′ + B ′) / 8
= (7.0 ・ 255) / 8
= 255/8
= 31.875 (19)
(See step S15).

  As described above, when the value of the pixel A ′ is “0”, the value of the pixel B ′ is “255”, and the value of the pixel C ′ is “255”, the encoding target pixel It is considered that the value of D is often close to “0”. Therefore, as shown in the equation (19), the weight of the pixel A ′ is made larger than the pixel B ′, and the prediction value P is set to a value close to “0”. It is considered that the difference can be reduced. This makes it possible to increase the compression rate while maintaining high image quality.

  In the equations (6) and (19), the reason why the weight of the pixel A ′ is not 100% is as follows. When the value of the pixel A ′ is “0”, the value of the pixel B ′ is “255”, and the value of the pixel C ′ is “255”, the value of the encoding target pixel D is “0”. It is thought that there are many cases close to. However, even in such a case, it is not impossible that the value of the encoding target pixel D is not close to “0”. In view of this, the weight of the pixel A ′ is not set to 100%.

  Next, FIG. 5C will be described. In FIG. 5C, the value of the pixel A ′ is “141”, the value of the pixel B ′ is “78”, and the value of the pixel C ′ is “0”. That is, the strong edge portion does not exist between the pixels A ′, B ′, and C ′.

When such an image is input to the image encoding device 10, the absolute value dH of the difference between the pixel C ′ and the pixel B ′ is expressed by the following equation (2):
dH = | C′−B ′ |
= | 0-78 |
= | -78 |
= 78 (20)
(See step S11).

Further, the absolute value dV of the difference between the pixel C ′ and the pixel A ′ is expressed by the following equation (3):
dV = | C′−A ′ |
= | 0-141 |
= | -141 |
= 141 (21)
(See step S12).

And the difference dF between the value dH and the value dV is given by the equation (4):
dF = dH−dV
= 78-141
= −63 (22)
(See step S13).

  Since the difference dF is smaller than the threshold Th (here, 64) and larger than the threshold -Th (here, -64) (see step S16), the weighting coefficient representing the weighting of the pixel A ′ and the pixel B ′ with respect to the encoding target pixel D m is calculated to be approximately −0.369140625 by linear interpolation (step S18).

Therefore, the predicted value P of the encoding target pixel D is expressed by the following equation (7):
P = 4/8 (A ′ + B ′) + m (B′−A ′)
= 4/8 (141 + 78) -0.369140625 (78-141)
= 109.5 + 23.255859375
= 132.755859375 (23)
(See step S19).

  Here, since dF <0, m <0, that is, the weight of the pixel B ′ is smaller than the weight of the pixel A ′.

  As described above, according to the present embodiment, there is a strong edge portion with a very simple configuration and processing in which only three pixels located at the upper left, upper, and left of the encoding target pixel are referred to. Even in such a case, an accurate predicted value can be calculated. Also, since no code is assigned to the identification of the predicted value, it is possible to perform fixed-length encoding.

Next, an image decoding apparatus according to an embodiment of the present invention will be described.
FIG. 6 shows a configuration example of an image decoding apparatus according to an embodiment of the present invention. The image decoding device 40 includes an image decoding unit 41 to which decoding target pixels encoded by the image encoding device 10 described above are sequentially input, and a line buffer (FIFO buffer) 42. The image decoding device 40 may be realized as a semiconductor integrated circuit.

  The image decoding unit 41 includes a first difference absolute value calculation unit 51, a second difference absolute value calculation unit 52, a difference calculation unit 53, a predicted value calculation unit 54, an inverse quantization unit 55, and a decoding unit. 56.

  The first difference absolute value calculation unit 51 calculates the absolute value of the difference between the pixel located at the upper left of the decoding target pixel and the pixel located above the decoding target pixel.

  The second difference absolute value calculation unit 52 calculates the absolute value of the difference between the pixel located on the upper left side of the decoding target pixel and the pixel located on the left side of the decoding target pixel.

The difference calculation unit 53 calculates the difference between the value calculated by the first difference absolute value calculation unit 51 and the value calculated by the second difference absolute value calculation unit 52.
The predicted value calculation unit 54 calculates the predicted value of the decoding target pixel based on the value calculated by the difference calculation unit 53.

The inverse quantization unit 55 inversely quantizes the decoding target pixel.
The decoding unit 56 decodes the decoding target pixel by adding the value inversely quantized by the inverse quantization unit 55 and the prediction value calculated by the prediction value calculation unit 54. The pixel decoded by the decoding unit 56 is output to the external circuit at the subsequent stage and stored in the line buffer 42, and the subsequent pixel (the pixel located right next to the current decoding target pixel, the current decoding) Used to decode a pixel located below the pixel to be converted, and a pixel located to the lower right of the current pixel to be decoded).

  Next, the operation of the image decoding device 40 will be described.

  FIG. 7 is a flowchart illustrating processing executed by the image decoding device 40. The image decoding device 40 executes the process shown in FIG. 7 every time a pixel (decoding target pixel) is input from the outside. When each pixel includes a plurality of components, the process shown in FIG. 7 may be executed for each component of the decoding target pixel.

First, the first difference absolute value calculation unit 51 decodes a pixel (here, referred to as pixel C ′ (see FIG. 2)) located at the upper left of the decoding target pixel (here, referred to as pixel D ′). The absolute value dH of the difference from a pixel (here, referred to as pixel B ′ (see FIG. 2)) located above the target pixel D ′ is dH = | C′−B ′ | (24)
(Step S31).

Next, the second absolute difference calculator 52 calculates a pixel C ′ located at the upper left of the decoding target pixel D ′ and a pixel located at the left of the decoding target pixel D ′ (here, the pixel A ′ ( The absolute value dV of the difference from (see FIG. 2))) is expressed as dV = | C′−A ′ | (25)
(Step S32).

Next, the difference calculation unit 53 sets the difference dF between the value dH calculated by the first difference absolute value calculation unit 51 and the value dV calculated by the second difference absolute value calculation unit 52 as dF = dH−dV.・ (26)
(Step S33).

  Next, the predicted value calculation unit 54 calculates the predicted value P of the decoding target pixel D ′ based on the value dF calculated by the difference calculation unit 53 (steps S34 to S39).

Specifically, when dF is larger than a given threshold Th (here, 64) (step S34), the predicted value calculation unit 54 sets the predicted value P to P = (A ′ + 7B ′) / 8. ... (27)
(Step S35).

Moreover, the predicted value calculation unit 54 sets the predicted value P to P = (7A ′ + B ′) / 8 when dF is smaller than a given threshold value −Th (here, −64) (step S36). ... (28)
(Step S37).

Further, when dF is larger than the threshold value -Th and smaller than the threshold value Th, the predicted value calculation unit 54 calculates a weighting coefficient m representing the weighting of the pixel A 'and the pixel B' with respect to the decoding target pixel D '( Step S38). For example, as shown in FIG. 4 described above, m = 3/8 when dF = Th, m = −3 / 8 when dF = −Th, and the value of dF between these two points. Depending on, linear interpolation may be performed. Then, the predicted value calculation unit 54 uses the weighting coefficient m calculated in step S38 to change the predicted value P to P = 4/8 (A ′ + B ′) + m (B′−A ′) (29) )
(Step S39).

  In Steps S34 to S35, when dF is larger than the threshold Th, calculating the predicted value P by Expression (27) can be considered as calculating the weighting coefficient m as 3/8. In steps S36 to S37, when dF is smaller than the threshold value -Th, calculating the predicted value P by equation (28) may be considered as calculating the weighting coefficient m to -3/8. it can.

Next, the inverse quantization unit 55 converts the value iCode of the decoding target pixel D ′ to d ′ = deQuant (iCode) (30)
Then, inverse quantization is performed (step S40). Here, deQuant () represents an inverse quantization operation. As the inverse quantization operation, various known operations can be used. For example, table conversion using an inverse quantization table can also be used.

Then, based on the predicted value P calculated by the predicted value calculation unit 54 and the value d ′ calculated by the inverse quantization unit 55, the decoding unit 56 sets the decoding target pixel D ′ to D ′ = P + d ′. ... (31)
(Step S41). This pixel D ′ is output to the external circuit at the subsequent stage and stored in the line buffer 42, and the subsequent pixel (the pixel located on the right side of the current decoding target pixel, the lower side of the current decoding target pixel) And the pixel located at the lower right of the current decoding target pixel).

Next, a system using the image encoding device and the image decoding device according to the present embodiment will be described.
FIG. 8 shows a configuration example of a system using an image encoding device and an image decoding device according to an embodiment of the present invention. The system 60 includes a semiconductor integrated circuit 70, an LCD controller 80, an LCD driver 90, and an LCD panel 100. The semiconductor integrated circuit 70 includes an image encoding device 10 according to an embodiment of the present invention, a VRAM 71, and an image decoding device 40 according to an embodiment of the present invention.

  The image encoding device 10 encodes the input image signal and outputs it to the VRAM 71. The VRAM 71 stores the image signal encoded by the image encoding device 10. The image decoding device 40 decodes the image signal stored in the VRAM 71 and outputs it to the LCD controller 80. The LCD controller 80 controls the LCD driver 90 based on the image signal decoded by the image decoding device 40. The LCD driver 90 drives the LCD panel 100 under the control of the LCD controller 80, and the LCD panel 100 is , Display an image.

  According to this system 60, since the input image signal is encoded and stored in the VRAM 71, the capacity of the VRAM 71 can be kept small.

  Although the present embodiment has been described in detail as described above, it will be easily understood by those skilled in the art that many modifications can be made without departing from the novel matters and effects of the present invention. Accordingly, all such modifications are intended to be included in the scope of the present invention. For example, a term described at least once together with a different term having a broader meaning or the same meaning in the specification or the drawings can be replaced with the different term in any part of the specification or the drawings. Further, the configuration and operation of the image encoding device, the image decoding device, and the system are not limited to those described in this embodiment, and various modifications can be made.

1 is a configuration example of an image encoding device according to an embodiment of the present invention. An example of image data input to the image encoding device 10 of FIG. The flowchart of the image coding method which concerns on one Embodiment of this invention. The figure which shows the example of the relationship between the difference value dF and the weighting coefficient m. An example of image data input to the image encoding device 10 of FIG. The structural example of the image decoding apparatus which concerns on one Embodiment of this invention. The flowchart of the image decoding method which concerns on one Embodiment of this invention. The structural example of the system which concerns on one Embodiment of this invention. The example of the image data input into the conventional image coding apparatus.

Explanation of symbols

10 image encoding device, 11 image encoding unit, 12, 42 line buffer,
21, 51 first difference absolute value calculation unit, 22, 52 second difference absolute value calculation unit,
23, 53 Difference calculation unit, 24, 54 Prediction value calculation unit, 25 Pixel prediction value difference calculation unit,
26 quantization unit, 27, 55 inverse quantization unit, 28, 56 decoding unit,
40 image decoding device, 41 image decoding unit, 60 system,
70 semiconductor integrated circuit, 71 VRAM, 80 LCD controller,
90 LCD driver, 100 LCD panel

Claims (11)

The encoding target pixel in the image data including a plurality of pixels is encoded in the first direction adjacent to the encoding target pixel in the first direction and in the second direction intersecting the first direction. Encoding based on a second pixel adjacent to the target pixel and a third pixel adjacent to the first pixel in the second direction and adjacent to the second pixel in the first direction An image encoding device for
When the absolute value of the difference between the third pixel and the second pixel is larger than the absolute value of the difference between the third pixel and the first pixel, the second pixel is changed to the first pixel. If the absolute value of the difference between the third pixel and the second pixel is smaller than the absolute value of the difference between the third pixel and the first pixel with a weight greater than A prediction value of the encoding target pixel is calculated by performing a weighting operation on the first pixel and the second pixel with a weight smaller than that of the first pixel, and the encoding target pixel An image encoding device that encodes the encoding target pixel by quantizing a difference from the predicted value. The encoding target pixel in the image data including a plurality of pixels is encoded in the first direction adjacent to the encoding target pixel in the first direction and in the second direction intersecting the first direction. Encoding based on a second pixel adjacent to the target pixel and a third pixel adjacent to the first pixel in the second direction and adjacent to the second pixel in the first direction An image encoding method for
Calculating an absolute value of a difference between the third pixel and the second pixel (a);
Calculating an absolute value of a difference between the third pixel and the first pixel (b);
Calculating a weighting coefficient representing weighting of the first pixel and the second pixel with respect to the encoding target pixel, wherein an absolute value of a difference between the third pixel and the second pixel is When the absolute value of the difference between the third pixel and the first pixel is larger than the absolute value of the difference, the weighting coefficient for calculating the second pixel as a weight greater than the first pixel is calculated, and the third pixel is calculated. When the absolute value of the difference between the second pixel and the second pixel is smaller than the absolute value of the difference between the third pixel and the first pixel, the second pixel is weighted smaller than the first pixel. (C) calculating the weighting coefficient
Calculating a predicted value of the encoding target pixel by performing a weighting operation on the first pixel and the second pixel using the weighting coefficient;
Calculating a difference between the encoding target pixel and the predicted value (e);
A step (f) of quantizing the value calculated in step (e);
An image encoding method including: In claim 2,
Step (c)
When the difference between the absolute value of the difference between the third pixel and the second pixel and the absolute value of the difference between the third pixel and the first pixel is greater than a given first threshold value Uses the given first value as the weighting coefficient, the absolute value of the difference between the third pixel and the second pixel, and the absolute value of the difference between the third pixel and the first pixel Is smaller than a given second threshold, the given second value is the weighting factor, and the absolute value of the difference between the third pixel and the second pixel is When the difference between the absolute value of the difference between the third pixel and the first pixel is smaller than the first threshold and larger than the second threshold, the first value and the second value An image encoding method characterized in that a value obtained by interpolation is used as the weighting coefficient. In claim 2 or 3,
Each of the plurality of pixels includes a plurality of components;
Step (a)-step (f) are performed about each of the component of the said encoding object pixel, The image coding method characterized by the above-mentioned. In claim 4,
Each of the plurality of pixels includes an R component, a G component, and a B component;
Step (a)-step (f) are performed about each of R component of the said encoding object pixel, G component, and B component, The image coding method characterized by the above-mentioned. In any of claims 2 to 5,
(G) dequantizing the value calculated in step (f);
Storing the value obtained by adding the predicted value calculated in step (d) and the value calculated in step (g) in a buffer (h);
Further including
The value stored in the buffer in step (h) is the fourth pixel adjacent to the encoding target pixel in the direction opposite to the first direction, and the encoding target pixel in the direction opposite to the second direction. And / or a fifth pixel adjacent to the fourth pixel in the direction opposite to the second direction and adjacent to the fifth pixel in the direction opposite to the first direction. An image encoding method, which is used when encoding a pixel. The first pixel adjacent to the decoding target pixel in the first direction is a decoding target pixel in the image data encoded by the image encoding method according to claim 2. A second pixel adjacent to the pixel to be decoded in a second direction intersecting the first direction, and adjacent to the first pixel in the second direction and in the first direction. An image decoding method for decoding based on a third pixel adjacent to the second pixel,
Calculating an absolute value of a difference between the third pixel and the second pixel (a);
Calculating an absolute value of a difference between the third pixel and the first pixel (b);
Calculating a weighting coefficient representing weighting of the first pixel and the second pixel with respect to the encoding target pixel, wherein an absolute value of a difference between the third pixel and the second pixel is When the absolute value of the difference between the third pixel and the first pixel is greater than the absolute value of the difference, the weighting coefficient for calculating the second pixel as a weight greater than that of the first pixel is calculated. When the absolute value of the difference between the second pixel and the second pixel is smaller than the absolute value of the difference between the third pixel and the first pixel, the second pixel is weighted smaller than the first pixel. (C) calculating the weighting coefficient
Calculating a predicted value of the decoding target pixel by performing a weighting operation on the first pixel and the second pixel using the weighting coefficient;
Dequantizing the decoding target pixel (e);
A step (f) of adding the predicted value calculated in step (d) and the value calculated in step (e);
An image decoding method including: The encoding target pixel in the image data including a plurality of pixels is encoded in the first direction adjacent to the encoding target pixel in the first direction and in the second direction intersecting the first direction. Encoding based on a second pixel adjacent to the target pixel and a third pixel adjacent to the first pixel in the second direction and adjacent to the second pixel in the first direction An image encoding device for
A first difference absolute value calculation unit for calculating an absolute value of a difference between the third pixel and the second pixel;
A second difference absolute value calculation unit for calculating an absolute value of a difference between the third pixel and the first pixel;
A difference calculation unit that calculates a difference between the absolute value of the difference calculated by the first difference absolute value calculation unit and the absolute value of the difference calculated by the second difference absolute value calculation unit;
A prediction value calculation unit that calculates a prediction value of the encoding target pixel based on the difference calculated by the difference calculation unit;
A pixel prediction value difference calculation unit that calculates a difference between the encoding target pixel and the prediction value calculated by the prediction value calculation unit;
A quantization unit that quantizes the difference calculated by the pixel prediction value difference calculation unit;
An image encoding device. In claim 8,
Each of the plurality of pixels includes a plurality of components;
A plurality of sets of the first difference absolute value calculation unit, the second difference absolute value calculation unit, the difference calculation unit, the prediction value calculation unit, the pixel prediction value difference calculation unit, and the quantization unit; An image encoding device, wherein each set encodes each component of the encoding target pixel in parallel. The decoding target pixel in the image data encoded by the image encoding device according to claim 8 or 9, the first pixel adjacent to the decoding target pixel in the first direction, and the first A second pixel that is adjacent to the pixel to be decoded in a second direction that intersects the direction of the second pixel, and that is adjacent to the first pixel in the second direction and the second pixel in the first direction. An image decoding device that performs decoding based on a third pixel adjacent to
A first difference absolute value calculation unit for calculating an absolute value of a difference between the third pixel and the second pixel;
A second difference absolute value calculation unit for calculating an absolute value of a difference between the third pixel and the first pixel;
A difference calculation unit that calculates a difference between the absolute value of the difference calculated by the first difference absolute value calculation unit and the absolute value of the difference calculated by the second difference absolute value calculation unit;
A prediction value calculation unit that calculates a prediction value of the decoding target pixel based on the difference calculated by the difference calculation unit;
An inverse quantization unit that inversely quantizes the decoding target pixel;
A decoding unit that decodes the decoding target pixel by adding the prediction value calculated by the prediction value calculation unit and the value calculated by the inverse quantization unit;
An image decoding apparatus. An image encoding device according to claim 8 or 9,
A memory for storing image data encoded by the image encoding device;
The image decoding device according to claim 10, which decodes image data stored in the memory;
A semiconductor integrated circuit.