KR20200104842A - Apparatus and method for encoding and decoding to image of ultra high definition resoutltion - Google Patents

Abstract

Provided are a device and method for encoding an ultra-high definition image using intra prediction encoding, and a device and method for decoding. The device for encoding may generate a prediction block of a current block according to any one of a plurality of predefined prediction block generation methods, and perform entropy encoding on at least one of encoding information and a difference block and output as a bitstream according to whether or not the difference block is encoded.

Description

An apparatus and method for encoding an ultra-high-resolution image, and a decoding apparatus and method TECHNICAL FIELD

The present invention relates to a technique for reducing computational complexity and improving encoding efficiency during intra prediction encoding in ultra-high resolution images.

Broadcasting services with high definition (HD) resolutions such as 1280×1024 and 1920×1080 are expanding not only in Korea but also worldwide. Accordingly, users are demanding higher resolution and higher quality content such as 4K (3840×2160) and 8K (7680×4320) in order to view more realistic images.

The number of pixels of a 4K image is four times the size of an HD image, and the number of pixels of an 8K (7680x4320) image is 16 times that of an HD image. Accordingly, 4K and 8K images can express more delicate and natural images than images with HD resolution. In general, H.264/AVC (MPEG-4 Part 10 Advanced Video Coding) is used to encode 4K or 8K video.

As such, as the resolution of the image increases, the spatial correlation between pixels may be greatly increased.

1 is a diagram comparing the number of pixels used to represent a part of an image for each resolution.

In FIG. 1, when the number of pixels used to express a part of an image at 480 × 270 resolution higher than QVGA (352 × 288) resolution is 4, a part of image is expressed in 960 × 540 resolution and FULL HDTV (1920 × 1080) The number of pixels used for this is 9 and 25, respectively. In addition, in the 4K (3840 × 2160) resolution, the number of pixels is 100, and in the 8K (7680 × 4320) resolution, a part of the image is represented by the number of 361 pixels. As such, images of 4K and 8K resolutions have an excessively large number of pixels used to represent a part of a specific image. In other words, images of 4K and 8K resolution have a much higher correlation between pixels than images of HD resolution or lower. In particular, in an image or a portion of an image that is not complex and has a flat characteristic, the correlation between pixels is further increased.

As described above, when intra-prediction is performed on an ultra-high resolution image having a very high correlation between pixels, such as 4K or 8K, pixel values of the current block can be almost predicted using only pixel values of neighboring blocks. Accordingly, when the correlation between pixels is high, the process of encoding and decoding the difference block increases computational complexity and decreases encoding efficiency.

Accordingly, when encoding and decoding an ultra-high resolution image having a high correlation between pixels, there is a need for an encoding and decoding technology capable of improving encoding efficiency while reducing computational complexity.

The present invention provides an encoding apparatus and method, and a decoding apparatus and method capable of reducing computational complexity according to intra prediction encoding for an ultra-high resolution image having high correlation between pixels.

In addition, the present invention provides an encoding apparatus and method, and a decoding apparatus and method capable of improving encoding efficiency by performing encoding and decoding using a prediction block having the best encoding efficiency among a plurality of prediction blocks.

An encoding apparatus according to an embodiment of the present invention includes an intra prediction unit that generates a prediction block of a current block based on a reference image, and a subtraction unit that generates a difference block through a difference between the current block and the generated prediction block. An encoding information generator for generating encoding information indicating whether the difference block is encoded based on whether the difference block is encoded, and entropy encoding at least one of the difference block and the encoding information based on the encoding information. It may include an entropy encoding unit.

Further, when the encoding information includes information indicating that the difference block is not encoded, the entropy encoding unit may generate a bitstream by encoding only the encoding information.

In addition, the intra prediction unit may generate the prediction block using any one of a plurality of prediction block generation methods.

In addition, the intra prediction unit may generate a plurality of prediction blocks of the current block according to a method of generating a plurality of prediction blocks.

In addition, the intra prediction unit may generate one or more prediction blocks of the current block based on a prediction block generation method used when generating prediction blocks of each of the reconstructed neighboring blocks.

In addition, it may include a prediction block selection unit that selects any one of the generated prediction blocks of the plurality of current blocks.

An encoding method according to an embodiment of the present invention includes generating a prediction block of a current block based on a reference image, generating a difference block through a difference between the current block and the generated prediction block, and the difference block Generating encoding information indicating whether to encode the difference block based on whether or not to encode the difference block, transforming and quantizing the difference block based on the encoding information, and the difference block and the difference block based on the encoding information It may include entropy encoding at least one of the encoding information.

Further, in the entropy encoding step, since the encoding information includes information indicating that the difference block is not encoded, only the encoding information may be encoded to generate a bitstream.

In addition, in the generating of the prediction block, a plurality of prediction blocks of the current block may be generated according to a method of generating a plurality of prediction blocks.

In addition, in the generating of the prediction block, one or more prediction blocks of the current block may be generated based on a prediction block generation method used when generating prediction blocks of each of the reconstructed neighboring blocks.

In addition, the method may further include selecting one of the prediction blocks of the generated current block having the best coding efficiency.

A decoding apparatus according to an embodiment of the present invention includes an entropy decoder that extracts encoded encoding information from a bitstream and decodes the encoded encoding information, and a screen for generating a prediction block of a current block using reconstructed neighboring blocks. An intra prediction unit, a difference block reconstructed based on the decoded encoding information, and an adder reconstructing a current block using at least one of a prediction block of the generated current block.

In addition, since the encoding information includes information indicating that the difference block is not encoded, the adder may reconstruct the current block using only the prediction block of the current block.

A decoding method according to an embodiment of the present invention includes extracting at least one of encoding information and prediction block selection information encoded from a bitstream, decoding the extracted encoding information and prediction block selection information, and reconstructed surroundings. Generating prediction blocks of a current block using a block, selecting any one of prediction blocks of the generated current block based on the decoded prediction block selection information, and based on the decoded encoding information And reconstructing the current block using at least one of the reconstructed difference block and the prediction block of the selected current block.

According to the present invention, calculation complexity can be reduced by outputting a bitstream using only a prediction block of a current block when intra prediction encoding an ultra-high resolution image having high correlation between pixels.

According to the present invention, encoding and decoding are performed using a prediction block having the best encoding efficiency among a plurality of prediction blocks, thereby improving encoding efficiency.

1 is a diagram comparing the number of pixels used to represent a part of an image for each resolution.
2 is a diagram illustrating a configuration of an encoding apparatus according to an embodiment of the present invention.
3 is a diagram illustrating an overall configuration of an encoding apparatus according to an embodiment of the present invention.
4 is a flowchart provided to explain the operation of the encoding apparatus according to an embodiment of the present invention.
5 is a block diagram showing the configuration of a decoding apparatus according to an embodiment of the present invention.
6 is a diagram showing the overall configuration of an encoding apparatus according to an embodiment of the present invention.
7 is a flowchart provided to explain the operation of the decoding apparatus according to an embodiment of the present invention.
8 is a diagram showing the configuration of an encoding apparatus according to an embodiment of the present invention.
9 is a flowchart provided to explain the operation of the encoding apparatus according to an embodiment of the present invention.
10 is a block diagram showing the configuration of a decoding apparatus according to an embodiment of the present invention.
11 is a flowchart provided to explain the operation of the decoding apparatus according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. However, the present invention is not limited or limited by the embodiments. In addition, the same reference numerals shown in each drawing denote the same member.

2 is a diagram illustrating a configuration of an encoding apparatus according to an embodiment of the present invention.

In FIG. 2, an input image may be processed in coding units having a size of 16×16 pixels in width and height. In this case, the encoding device may output a bitstream generated through intra prediction encoding or inter prediction encoding. First, the intra prediction encoding process will be described with reference to FIG. 2, and the inter prediction encoding process will be described later with reference to FIG. 3.

Referring to FIG. 2, the encoding apparatus 200 includes an intra prediction unit 210, a reference image buffer 220, a subtraction unit 230, an encoding information generation unit 240, a transformation and quantization unit 250, and entropy encoding. A unit 260, an inverse quantization and inverse transform unit 270, and an addition unit 280 may be included.

The intra prediction unit 210 may generate a prediction block of the current block based on the reference image stored in the reference image buffer 220. In this case, when using a neighboring block reconstructed as a reference image, the intra prediction unit 210 uses one of a plurality of predefined prediction block generation methods and pixel values of neighboring blocks spatially adjacent to the current block. A prediction block of blocks can be generated.

For example, (1) the intra prediction unit 210 may generate a prediction block of the current block by copying (padding) pixels of a neighboring block adjacent to the current block as it is. In this case, the pixel copied (padded) to the current block may vary according to the predicted direction, and the predicted direction information may be encoded and included in the bitstream. In other words, the intra prediction unit 210 may generate a prediction block of the current block by using the intra prediction method of H.264/AVC.

As another example, (2) the intra prediction unit 210 determines a prediction direction based on the similarity between the current block and neighboring blocks, and copies pixels of the neighboring blocks according to the determined prediction direction to generate a prediction block of the current block. I can. In other words, the intra prediction unit 210 selects a neighboring block located in the determined prediction direction from among a plurality of neighboring blocks adjacent to the current block, and copies pixels of the selected neighboring block to generate a prediction block of the current block. I can.

As another example, (3) the intra-prediction unit 210 may extract a block most similar to the current block through matching in a current screen region in which encoding has already been performed. In addition, the intra prediction unit 210 may generate a prediction block of the current block by using the extracted block. In this case, difference information about the position between the current block and the prediction block of the current block may be included in the bitstream.

As another example, (4) the intra prediction unit 210 may find a shape most similar to a neighboring block in a current screen area in which encoding has already been performed, and generate a block surrounded by the found shape as a prediction block of the current block.

As another example, (5) the intra prediction unit 210 may generate a prediction block of the current block by mixing a plurality of predefined prediction block generation methods. In this case, the intra prediction unit 210 may generate a prediction block by using an average value of prediction blocks generated using a plurality of prediction block generation methods. In addition, a prediction block may be generated using the sum of weights according to each of the plurality of prediction block generation methods.

The subtraction unit 230 may generate a residual block by subtracting the current block and the prediction block of the current block.

The encoding information generator 240 may generate encoding information based on whether the difference block is encoded. For example, a similarity between a current block and a prediction block of the current block may be used to determine whether to encode or not encode the difference block.

When the similarity between the current block and the prediction block of the current block is used, the encoding information generation unit 240 uses SAD (Sum of Absolute Differences), SATD (Sum of Absolute Transformed Differences), SSD (sum of squared difference), etc. Thus, the similarity between the current block and the prediction block of the current block can be calculated. In addition, the encoding information generator 240 may generate encoding information based on the calculated similarity. Here, the encoding information may include information indicating whether to encode the difference block or not.

For example, when the similarity between the current block and the prediction block of the current block is greater than or equal to a preset reference similarity, the encoding information generator 240 may generate encoding information including information indicating that the difference block is not encoded.

As another example, when the similarity between the current block and the prediction block of the current block is less than the reference similarity, the encoding information generator 240 may generate encoding information including information indicating that the difference block is encoded.

Then, the transform and quantization unit 250 may transform and quantize the difference block based on the encoding information. In addition, the entropy encoder 260 may generate a bitstream by entropy encoding at least one of the encoding information and the difference block based on the encoding information.

For example, when the encoding information includes information indicating that the difference block is not encoded, the transform and quantization unit 250 may not transform and quantize the difference block. Then, the entropy encoder 260 may generate a bitstream by entropy encoding only the encoded information without a difference block. Accordingly, computational complexity and encoding time due to transforming and quantizing the difference block may be reduced. As such, when the encoding information includes information indicating that the difference block is not encoded, the generated bitstream may include encoded encoding information.

As another example, when the encoding information includes information indicating that the difference block is encoded, the transform and quantization unit 250 may transform and quantize the difference block. Then, the entropy encoding unit 260 may entropy-encode the transformed and quantized difference block and entropy-encode the encoding information. In addition, the entropy encoder 260 may generate a bitstream by mixing the encoded difference block and the encoded encoding information. As described above, when the encoding information includes information indicating that the difference block is encoded, the generated bitstream may include both the encoded difference block and the encoded encoding information.

In this case, the transform and quantization unit 250 may perform different transform and quantization on the difference block according to the intra prediction coding mode.

For example, in the case of using 16×16 intra prediction coding, the transform and quantization unit 250 transforms the difference block to generate a transform coefficient, collects only the DC coefficient from the generated transform coefficients, and transforms the Hadamard again. Can be done. Then, the entropy encoder 260 may entropy-encode only the Hadamard-transformed DC coefficients and output them as a bitstream.

As another example, when 8×8 and 4×4 intra prediction coding modes other than 16×16 are used, the transform and quantization unit 250 transforms the difference block to generate a transform coefficient, and the generated transform A quantized coefficient may be generated by quantizing a coefficient according to a quantization parameter. Then, the entropy encoding unit 260 may entropy-encode the generated quantization coefficient according to a probability distribution and output it as a bitstream.

The inverse quantization and inverse transform unit 270 may inverse quantize and inverse transform the quantization coefficient output from the transform and quantization unit 250 to decode the difference block.

In addition, the addition unit 280 may restore the current block by adding the decoded difference block and the prediction block of the current block. Then, the reference image buffer 220 may store the reconstructed current block.

In this way, as the restored current block is stored in the reference image buffer 220, the neighboring blocks restored in the previous step may be stored in the reference image buffer 220 as a reference image. Then, when the encoding information includes information indicating that the difference block is not encoded, the reconstructed neighboring block stored in the reference image buffer 220 is a reconstructed block using only the prediction block of the neighboring block. In addition, when the encoding information includes information indicating that the difference block is encoded, the reconstructed neighboring block stored in the reference image buffer 220 is a reconstructed block using the difference block and the prediction block of the neighboring block. Accordingly, the intra-prediction unit 210 uses the neighboring block reconstructed using only the prediction block of the neighboring block according to the encoding information, or the neighboring block reconstructed using the reconstructed difference block and the predicted block of the neighboring block. A prediction block of blocks can be generated.

In FIG. 2 above, a configuration for performing intra prediction encoding on an input image has been described for convenience of description, but hereinafter, a configuration for performing both intra prediction encoding and inter prediction encoding on an input image will be described with reference to FIG. 3. I will do it.

3 is a diagram illustrating an overall configuration of an encoding apparatus according to an embodiment of the present invention.

In FIG. 3, components mentioned while describing the operation of FIG. 2 are substantially the same, and thus, a duplicate description will be omitted. The encoding apparatus 300 of FIG. 3 may further include a motion prediction unit 310, a motion compensation unit 320, and a deblocking unit 350 in the encoding apparatus 200 of FIG. 2.

The motion prediction unit 310 may calculate a motion vector by finding an area that best matches the current block among reference images using various motion estimation algorithms such as BMA (Block Matching Algorithm), Phase Correlation, HSBMA, etc. have

Then, the motion compensation unit 320 may perform motion compensation using the calculated motion vector to generate a prediction block of the current block. As such, the encoding apparatus according to an embodiment of the present invention may perform inter prediction encoding through motion prediction and compensation. Then, the subtraction unit 330 may generate a difference block by subtracting the prediction block and the current block of the current block generated through inter prediction encoding or intra prediction encoding.

The adder 340 may reconstruct the current block by using at least one of the prediction block and the reconstructed difference block of the current block. In addition, the reconstructed difference block may be stored in the reference image buffer.

For example, when the difference block is not transformed and quantized by the transform and quantization unit, the adder 304 may reconstruct the current block only with the prediction block of the current block. As another example, when the difference block is transformed and quantized in the transform and quantization unit, the addition unit 304 may reconstruct the current block by adding the prediction block of the current block and the reconstructed difference block.

Then, the deblocking unit 340 may remove a blocking artifact from the restored current block using the deblocking filter and store it in the reference image buffer.

In FIG. 3, inter prediction encoding is divided into 16×16, 16×8, 8×16, and 8×8 inter prediction encoding modes, and 8×8 inter prediction encoding modes are again 8×8 and 8×4. , 4×8, and 4×4 sub-pictures can be divided into prediction encoding modes.

4 is a flowchart provided to explain the operation of the encoding apparatus according to an embodiment of the present invention.

First, in step 410, the intra prediction unit 210 may generate a prediction block of a current block based on a reference image.

For example, the intra prediction unit 210 may generate a prediction block of the current block using any one of a plurality of predefined prediction block generation methods. Here, the reconstructed neighboring block may be used as the reference image. In this case, the reconstructed neighboring block may be reconstructed using only the prediction block of the neighboring block, or may be a block reconstructed using the prediction block of the neighboring block and the decoded difference block. Then, the intra prediction unit 210 may generate a prediction block of the current block by using the reconstructed neighboring blocks.

Subsequently, in step 420, the subtraction unit 230 may generate a difference block by subtracting the current block and the prediction block of the current block.

Then, in step 430, the encoding information generator 240 may generate encoding information based on whether or not the difference block is encoded. In this case, whether to encode the difference block may be determined based on a similarity between the current block and the prediction block of the current block.

For example, in the case of using the similarity, the encoding information generation unit 240 uses sum of absolute differences (SAD), sum of absolute transformed differences (SATD), sum of squared difference (SSD), etc. Similarity between prediction blocks can be calculated. In addition, the encoding information generator 240 may generate encoding information based on the calculated similarity.

In this case, when the calculated similarity is greater than or equal to the reference similarity, the encoding information generator 240 may generate encoding information to include information indicating that the difference block is not encoded. In addition, when the calculated similarity is less than the reference similarity, the encoding information generator 240 may generate encoding information to include information indicating that the difference block is encoded.

Subsequently, in step 440, the transform and quantization unit 250 may transform and quantize the difference block based on the encoding information.

In this case, when the encoding information includes information indicating that the difference block is to be encoded, the transform and quantization unit 250 may transform and quantize the difference block.

In addition, when the encoding information includes information indicating that the difference block is not encoded, the transform and quantization unit 250 may not transform and quantize the difference block.

Then, in step 450, the entropy encoding unit 260 may entropy-encode at least one of the transformed and quantized difference blocks and encoding information and output them as a bitstream.

For example, when a difference block is transformed and quantized based on encoding information, the entropy encoder 260 may entropy-encode the transformed and quantized difference block and entropy-encode the encoded information. In addition, the entropy encoder 260 may transmit a bitstream generated by mixing the encoded difference block and encoding information to the decoding apparatus.

As another example, when the difference block is not transformed and quantized based on the encoding information, the entropy encoder 260 may entropy-encode only the encoding information without the difference block and output it as a bitstream.

5 is a block diagram showing the configuration of a decoding apparatus according to an embodiment of the present invention.

Referring to FIG. 5, the decoding apparatus 500 includes an intra prediction unit 510, a reference image buffer 520, an entropy decoding unit 530, an inverse quantization and inverse transform unit 540, and an adder 550. can do.

The intra prediction unit 510 may generate a prediction block of the current block based on the reference image. In this case, the intra prediction unit 520 may generate a prediction block of the current block using any one of a plurality of predefined prediction block generation methods. Here, since the predefined method of generating a plurality of prediction blocks has already been described by the intra prediction unit 210 of FIG. 2, a redundant description will be omitted.

As an example, the intra prediction unit 510 may generate a prediction block of the current block by using the reconstructed neighboring blocks stored in the reference image buffer 520. In this case, depending on whether the difference block is encoded, the reconstructed neighboring block may be reconstructed using only the prediction block of the neighboring block, or a reconstructed block using the prediction block of the neighboring block and the decoded difference block.

The entropy decoder 530 may demultiplex the bitstream to extract encoded encoding information, and entropy-decode the encoded encoding information.

In this case, when the encoding information includes information indicating that the difference block is encoded, the entropy decoder 530 may extract not only the encoded encoding information but also the encoded difference block while demultiplexing the bitstream. In addition, the entropy decoder 530 may decode the encoded encoding information and decode the encoded difference block.

Further, when the encoding information includes information indicating that the difference block is not encoded, the entropy decoder 530 may extract the encoded encoding information while demultiplexing the bitstream. That is, when the encoding information includes information indicating that the difference block is not encoded, the encoded difference block may not be included in the bitstream. Accordingly, the entropy decoder 530 may entropy-decode only the encoded information.

The inverse quantization and inverse transform unit 540 may decode the difference block by inverse quantizing and inverse transforming the difference block based on the decoded encoding information.

As an example, when the encoding information includes information indicating that the difference block is to be encoded, the inverse quantization and inverse transform unit 540 inverse quantizes the encoded difference block using variable length decoding according to a probability distribution, thereby calculating a quantization coefficient. Can be printed. In this case, the output quantization coefficient may include only DC coefficient information or both DC coefficient information and AC coefficient information.

As another example, when the encoding information includes information indicating that the difference block is not encoded, the inverse quantization and inverse transform unit 540 may not operate because there is no encoded difference block extracted from the bitstream. In other words, the inverse quantization and inverse transform unit 540 can operate only when an encoded difference block is included in the bitstream.

The adder 550 may reconstruct the current block by using at least one of a difference block decoded based on the encoding information and a prediction block of the current block.

For example, when the encoding information includes information indicating that the difference block is to be encoded, the adder 550 may reconstruct the current block by adding the reconstructed difference block and the prediction block of the current block.

As another example, when the encoding information includes information indicating that the difference block is not encoded, the adder 550 may reconstruct the current block using only the prediction block of the current block generated by the intra prediction unit 510. .

The reference image buffer 520 may store the reconstructed current block.

6 is a diagram showing the overall configuration of an encoding apparatus according to an embodiment of the present invention.

In FIG. 6, since the components mentioned while describing the operation of FIG. 5 are substantially the same, a redundant description will be omitted. The decoding apparatus 600 of FIG. 6 may further include a motion compensation unit 610 and a deblocking unit 620 in the encoding apparatus 500 of FIG. 5.

The motion compensation unit 610 may perform motion compensation on a reference image using a motion vector extracted from a bitstream through entropy decoding. In addition, the motion compensation unit 610 may generate a prediction block of the current block through motion compensation.

The deblocking unit 620 may output a reconstructed image by removing a blocking phenomenon from the reconstructed current block. In addition, the reconstructed image may be stored in the reference image buffer.

7 is a flowchart provided to explain the operation of the decoding apparatus according to an embodiment of the present invention.

First, in step 710, the intra prediction unit 510 may generate a prediction block of the current block using a reference image.

As an example, the intra prediction unit 510 may generate a prediction block of the current block by using the reconstructed neighboring blocks stored in the reference image buffer 520. In this case, the intra prediction unit 510 may generate a prediction block of the current block according to any one of a plurality of predefined prediction block generation methods, as already described in the intra prediction unit 210 of FIG. 2. have.

Subsequently, in step 720, the entropy decoder 530 may extract the encoded information by demultiplexing the bitstream.

Further, in step 730, the entropy decoder 530 may entropy-decode the encoded encoding information. Here, the encoding information may include information indicating that the difference block is encoded or information indicating that the difference block is not encoded.

In this case, when the encoding information includes information indicating that the difference block is to be encoded, the entropy decoder 530 may extract the encoded difference block from the bitstream and perform entropy decoding.

Subsequently, in step 740, the inverse quantization and inverse transform unit 540 may inverse quantize and inversely transform the difference block based on the decoded encoding information.

For example, when the encoding information includes information indicating that the difference block is to be encoded, the inverse quantization and inverse transform unit 540 may inverse quantize the encoded difference block and reconstruct the difference block by inverse transformation.

As another example, when the encoding information includes information indicating that the difference block is not encoded, the inverse quantization and inverse transform unit 540 may skip the operation because there is no difference block extracted from the bitstream. In other words, the inverse quantization and inverse transform unit 540 can operate only when an encoded difference block is included in the bitstream. Accordingly, computational complexity of the decoding apparatus may be reduced, and a decoding time may be shortened.

In addition, in step 750, the adder 550 may reconstruct the current block based on at least one of the prediction block and the reconstructed difference block of the current block.

For example, when the encoding information includes information indicating that the difference block is to be encoded, the adder 550 may reconstruct the current block by adding the prediction block of the current block and the reconstructed difference block. Then, the current block, which is the image of the present application, may be displayed.

As another example, when the encoding information includes information indicating that the difference block is not encoded, the adder 550 may reconstruct the current block using only the prediction block of the current block. In other words, the reconstructed image may consist of only the prediction block of the current block.

In the above, a configuration for generating a prediction block of a current block by using any one of a plurality of predefined prediction block generations has been described with reference to FIGS. 2 to 7. In particular, a configuration in which the encoding and decoding apparatus according to an embodiment of the present invention reconstructs the current block using only the prediction block of the current block based on whether or not the difference block is encoded has been described.

Hereinafter, a plurality of prediction blocks of the current block are generated according to a prediction block generation method used when generating a prediction block of a neighboring block or a plurality of predefined prediction block generation methods, and any one of the prediction blocks of the current block is selected. Hereinafter, a configuration for encoding and decoding will be described.

8 is a diagram showing the configuration of an encoding apparatus according to an embodiment of the present invention.

FIG. 8 further includes a prediction block selection unit in the encoding apparatus of FIG. 2, and an operation of the intra prediction unit is different from the intra prediction unit of FIG. 2. Accordingly, among the constituent elements of FIG. 8, names overlapping with FIG. 2 will be omitted, and description will be made focusing on the prediction block selection unit and the intra prediction unit.

Referring to FIG. 8, the encoding apparatus 800 includes an intra prediction unit 810, a reference image buffer 820, a prediction block selection unit 830, a subtraction unit 840, an encoding information generation unit 850, a transform and A quantization unit 860, an entropy encoding unit 870, an inverse quantization and inverse transform unit 880, and an adder 890 may be included.

The intra prediction unit 810 may generate one or more prediction blocks of the current block by using the reference image stored in the reference image buffer 820. In this case, the intra-prediction unit 810 is used to generate a plurality of predefined prediction blocks or to generate prediction blocks of each of the reconstructed neighboring blocks based on the number of prediction block generation methods used when generating prediction blocks of neighboring blocks. A prediction block of the current block may be generated according to the predicted block generation method.

For example, if the prediction block generation methods used to generate the prediction blocks of the neighboring blocks are all the same, the intra prediction unit 810 may generate the prediction blocks of the current block according to the prediction block generation method of the neighboring blocks using the same. I can. In other words, when prediction blocks of neighboring blocks are generated according to intra prediction 1, the intra prediction unit 810 may generate a prediction block of the current block using the reference image according to intra prediction 1. Then, since the prediction block of the generated current block is one, the prediction block selector 830 may output the prediction block of the generated current block as it is.

As another example, when two or less prediction block generation methods are used when generating prediction blocks of neighboring blocks, the intra-prediction unit 810 may generate two current prediction blocks according to the two prediction block generation methods used when generating prediction blocks of neighboring blocks. A prediction block of blocks can be generated.

Then, the prediction block selector 830 may select a prediction block having good coding efficiency from among prediction blocks of the two current blocks. For example, the prediction block selection unit 830 may select a prediction block having good coding efficiency by using the similarity between the prediction block of the current block and the current block, rate-distortion cost, and the like. In addition, the prediction block selection unit 830 may generate prediction block selection information including information indicating a prediction block generation method used when generating a prediction block of the selected current block. In this case, the prediction block selection unit 830 may calculate the similarity using SAD, SSD, SATD, or the like.

As another example, when three or more prediction block generation methods are used to generate prediction blocks of neighboring blocks, the intra prediction unit 810 may generate a plurality of prediction blocks of the current block according to a plurality of predefined prediction block generation methods. Can produce dogs. For example, when the prediction block generation methods of (1) to (5) described in the intra prediction unit 210 of FIG. 2 are predefined, the intra prediction unit 810 generates five predefined prediction blocks. Five prediction blocks of the current block can be generated using methods. In other words, in the case of three or more, the intra prediction unit 810 may generate the prediction blocks of the current block using a plurality of predefined prediction block generation methods instead of the prediction block generation method used when generating the prediction blocks of neighboring blocks. I can.

Then, the prediction block selection unit 830 may select one of the prediction blocks of the current block with the best coding efficiency. In addition, the prediction block selection unit 830 may generate prediction block selection information including information indicating a prediction block generation method used when generating a prediction block of the selected current block.

The addition unit 840 may generate a difference block by subtracting the prediction block and the current block of the selected current block.

The encoding information generator 850 may generate encoding information based on whether the difference block is encoded. For example, a similarity between a current block and a prediction block of the current block may be used to determine whether to encode or not encode the difference block. Here, the encoding information may include information indicating whether to encode the difference block or not.

The transform and quantization unit 860 may transform and quantize the difference block based on the encoding information.

The entropy encoder 870 may entropy-encode at least one of encoding information, prediction block selection information, and difference block and output the encoded information as a bitstream. For example, when the difference block is not encoded, the entropy encoder 870 may entropy-encode the encoding information and the prediction block selection information and output the encoded information as a bitstream. In addition, when the difference block is encoded, the encoding information, prediction block selection information, and the difference block may be entropy-encoded and output as a bitstream.

The inverse quantization and inverse transform unit 880 may restore the difference block by inverse quantization and inverse transform of the transformed and quantized difference block.

The adder 890 may reconstruct the current block by adding the difference block reconstructed based on the encoding information and the prediction block of the selected current block. Then, the reference image buffer 820 may store the reconstructed current block. In this case, when the difference block is not encoded, the adder 890 may reconstruct the current block only with the prediction block of the selected current block.

9 is a flowchart provided to explain the operation of the encoding apparatus according to an embodiment of the present invention.

First, in step 910, the prediction block generator 810 may generate one or more prediction blocks of the current block using a reference image.

For example, the intra prediction unit 810 may generate a prediction block of the current block according to a prediction block generation method used when generating prediction blocks of each of the reconstructed neighboring blocks.

As another example, the intra prediction unit 810 may generate a plurality of prediction blocks of the current block according to a predefined plurality of prediction block generation methods based on the number of prediction block generation methods used when generating prediction blocks of neighboring blocks. I can.

Subsequently, in step 920, the prediction block selection unit 820 may select any one of the prediction blocks of the current block as prediction blocks of the plurality of current blocks are generated. In this case, the prediction block selection unit 820 may select a prediction block having the best encoding efficiency from among prediction blocks of a plurality of current blocks.

For example, when using the similarity between the current block and the prediction block of the generated current block, the prediction block selection unit 820 may determine that the higher the similarity, the better the encoding efficiency. Accordingly, the prediction block selection unit 820 may select a prediction block having the highest similarity among prediction blocks of the current block.

As another example, the prediction block selection unit 820 may select a prediction block having the best coding efficiency using rate-distortion cost.

Then, in step 930, the subtraction unit 840 may generate a difference block by subtracting the prediction block and the current block of the selected current block.

Subsequently, in step 940, the encoding information generator 850 may generate encoding information based on whether the difference block is encoded.

In step 950, the transform and quantization unit 860 may transform and quantize the difference block based on the encoding information.

Subsequently, in step 960, the entropy encoder 870 may entropy-encode at least one of encoding information, prediction block selection information, and difference block based on the encoding information.

For example, when the difference block is encoded, the entropy encoding unit 870 entropy-encodes each of the encoding information, prediction block selection information, and transformed and quantized difference blocks, and encoded encoding information, encoded prediction block selection information, and The encoded difference blocks may be mixed and output as a bitstream.

10 is a block diagram showing the configuration of a decoding apparatus according to an embodiment of the present invention.

The decoding apparatus 1000 of FIG. 10 may further include a prediction block selection unit 1050 in the encoding apparatus 400 of FIG. 4. Accordingly, components mentioned while describing the operation of FIG. 4 in FIG. 10 are substantially the same, and thus, a duplicate description will be omitted.

Referring to FIG. 10, the decoding apparatus 1000 includes an intra prediction unit 1010, a reference image buffer 1020, an entropy decoding unit 1030, an inverse quantization and inverse transform unit 1040, and an adder 1060. can do.

The intra prediction unit 1010 may generate a prediction block of the current block based on the reference image stored in the reference image buffer 1020.

For example, the intra prediction unit 810 may generate a prediction block of the current block according to a prediction block generation method used when generating prediction blocks of each of the reconstructed neighboring blocks.

As another example, the intra prediction unit 810 may generate a plurality of prediction blocks of the current block according to a predefined plurality of prediction block generation methods based on the number of prediction block generation methods used when generating prediction blocks of neighboring blocks. I can. Here, as a method of generating a plurality of prediction blocks, predefined prediction block generation methods of (1) to (5) described in the intra prediction unit 210 of FIG. 2 may be used.

The entropy decoder 1030 may demultiplex the bitstream to extract at least one of encoded encoding information, an encoded difference block, and encoded prediction block selection information. In addition, the entropy decoder 1030 may entropy-decode the encoded encoding information, the encoded difference block, and the encoded prediction block selection information.

For example, as the encoding information includes information indicating that the difference block is not encoded, the entropy decoder 1030 may entropy-decode only the encoded encoding information and the encoded prediction block selection information.

As another example, as the encoding information includes information indicating that the difference block is encoded, the entropy decoder 1030 may entropy-decode all of the encoded encoding information, the encoded difference block, and the encoded prediction block selection information. .

The prediction block selection unit 1050 may select any one of one or more prediction blocks generated based on prediction block selection information.

The inverse quantization and inverse transform unit 1040 may inverse quantize and inverse transform the encoded difference block based on the encoding information.

As an example, as the encoding information includes information indicating that the difference block is to be encoded, the inverse quantization and inverse transform unit 1040 may inverse quantize the encoded difference block and reconstruct the difference block by inverse transformation. In this case, since the encoding information includes information indicating that the difference block is not encoded, the inverse quantization and inverse transform unit 1040 may omit the operation.

The adder 1060 may reconstruct the current block by using at least one of the reconstructed difference block and the prediction block of the selected current block.

For example, when the difference block is reconstructed as the encoding information includes information indicating that the difference block is to be encoded, the adder 1060 may restore the current block by adding the reconstructed difference block and the prediction block of the selected current block. have.

As another example, when the inverse quantization and inverse transform units are omitted because the encoding information includes information indicating that the difference block is not encoded, the adder 1060 may reconstruct the current block only with the prediction block of the selected current block. . Then, a reconstructed image composed of only prediction blocks of the current block may be displayed.

In addition, the reference image buffer 1020 may store the restored current block.

11 is a flowchart provided to explain the operation of the decoding apparatus according to an embodiment of the present invention.

First, in step 1110, the intra prediction unit 1030 may generate one or more prediction blocks of the current block.

For example, the intra prediction unit 810 may generate a prediction block of the current block according to a prediction block generation method used when generating prediction blocks of each of the reconstructed neighboring blocks.

As another example, the intra prediction unit 810 may generate a plurality of prediction blocks of the current block according to a predefined plurality of prediction block generation methods based on the number of prediction block generation methods used when generating prediction blocks of neighboring blocks. I can.

Subsequently, in step 1120, the entropy decoder 1030 demultiplexes the bitstream to extract the encoded encoding information and the encoded prediction block selection information, and entropy-encode the encoded encoding information and the encoded prediction block selection information. have.

In this case, if the encoding information includes information indicating that the difference block is to be encoded, the entropy decoder 1030 may extract the encoded difference block from the bitstream through demultiplexing and perform entropy decoding.

Then, in step 1130, the inverse quantization and inverse transform unit 1040 may restore the difference block by inverse quantization and inverse transformation of the encoded difference block.

Subsequently, in step 1140, the prediction block selection unit 1050 may select any one of prediction blocks of one or more current blocks based on prediction block selection information.

Further, in step 1150, the adder 1060 may reconstruct the current block by using at least one of the prediction block and the reconstructed difference block of the selected current block. Then, the reference image buffer 1020 may store the reconstructed current block.

For example, when the operation of the inverse quantization and inverse transform unit 1040 is omitted because the encoding information includes information indicating that the difference block is not encoded, the addition unit 1060 determines the current block using only the prediction block of the selected current block. Can be restored. In other words, since the coded difference block is not included in the bitstream, the reconstructed difference block does not exist. Accordingly, the reconstructed image may be composed of only the prediction blocks of the current block selected based on the prediction block selection information.

As another example, when the difference block is reconstructed as the encoding information includes information indicating that the difference block is to be encoded, the addition unit 1060 may restore the current block by adding the prediction block of the selected current block and the reconstructed difference block. have.

As described above, although the present invention has been described by limited embodiments and drawings, the present invention is not limited to the above embodiments, and various modifications and variations from these descriptions for those of ordinary skill in the field to which the present invention pertains. This is possible.

Therefore, the scope of the present invention is limited to the described embodiments and should not be defined, but should be defined by the claims to be described later as well as equivalents to the claims.

210: intra-screen prediction unit
220: reference image buffer
230: subtraction
240: encoding information generation unit
250: transform and quantization unit
260: entropy encoding unit
270: Inverse quantization and inverse transform unit
280: adder

Claims (25)

Performing prediction on the current block; And
Generating a bitstream including selection information based on the prediction
Including,
The bitstream includes information on a transformed and quantized difference block generated based on a difference block that is a difference between the current block and the prediction block,
In decoding the current block using the bitstream, a prediction block for the current block is generated according to a prediction block generation method for the current block, and the prediction block generation method for the current block is It is determined based on methods of generating a plurality of prediction blocks of a plurality of neighboring blocks,
In the decoding of the current block, the plurality of neighboring blocks are blocks spatially adjacent to the current block,
In the decoding of the current block, the plurality of neighboring blocks are decoded blocks,
In the decoding of the current block, the methods of generating a plurality of prediction blocks of the plurality of neighboring blocks are prediction block generation methods respectively used when generating a plurality of prediction blocks of the plurality of neighboring blocks,
In decoding the current block, a method of generating a prediction block for the current block is determined based on the selection information. The method of claim 1,
The selection information indicates a prediction block generation method used to generate a prediction block of the current block. The method of claim 1,
The selection information indicates a prediction block generation method used to generate a prediction block of the current block among the plurality of prediction block generation methods. The method of claim 1,
In the decoding of the current block, as the selection information indicates one prediction block generation method among the plurality of prediction block generation methods, the prediction block of the current block is generated by the one prediction block generation method. Encoding method. The method of claim 1,
In decoding the current block, a method of generating a prediction block of a neighboring block of the current block specified by the selection information is used as a method of generating a prediction block for the current block. The method of claim 1,
In decoding the current block, a method of generating a prediction block for the current block is determined based on the number of different values for a method of generating a prediction block used for predictions of the neighboring blocks. The method of claim 1,
In decoding the current block, the prediction for the current block is intra prediction. The method of claim 1,
In decoding the current block, the prediction for the current block is a prediction using a specified decoded block as the prediction block for the current block. A computer-readable recording medium for storing the bitstream generated by the encoding method of claim 1. Extracting selection information from the bitstream;
Generating a prediction block by performing prediction on the current block; And
Generating a reconstructed block for a current block using the difference block and the prediction block
Including,
The prediction block is generated according to a prediction block generation method for the current block, and the prediction block generation method for the current block is determined based on a plurality of prediction block generation methods of a plurality of neighboring blocks of the current block,
In the decoding of the current block, the plurality of neighboring blocks are blocks spatially adjacent to the current block,
The plurality of neighboring blocks are decoded blocks,
The plurality of prediction block generation methods of the plurality of neighboring blocks are prediction block generation methods each used when generating a plurality of prediction blocks of the plurality of neighboring blocks,
A method of generating a prediction block for the current block is determined based on the selection information. The method of claim 10,
The selection information is a decoding method indicating a method of generating a prediction block used to generate the prediction block of the current block. The method of claim 10,
The selection information is a decoding method indicating a prediction block generation method used for generating the prediction block of the current block among the plurality of prediction block generation methods. The method of claim 10,
A decoding method in which the prediction block of the current block is generated by the one prediction block generation method as the selection information indicates one prediction block generation method among the plurality of prediction block generation methods. The method of claim 10,
A decoding method in which a prediction block generation method of a neighboring block of the current block specified by the selection information is used as a prediction block generation method of the current block. The method of claim 10,
The method of generating the prediction block for the current block is determined based on the number of different values for the prediction block generation method used for predictions of the neighboring blocks. The method of claim 10,
The prediction for the current block is intra prediction. The method of claim 10,
The prediction for the current block is a prediction using a specified decoded block as the prediction block for the current block. A computer-readable recording medium storing a bitstream for decoding an image, comprising:
The bitstream,
Information of transformed and quantized difference blocks; And
Optional information
Including,
In the decoding, a difference block is generated based on the transformed and quantized difference block, and a reconstructed block for the current block is generated based on the difference block and a prediction block for the current block, and the prediction block Is generated according to the prediction block generation method for the current block, and the prediction block generation method for the current block is determined based on a plurality of prediction block generation methods of a plurality of neighboring blocks of the current block,
In the decoding of the current block, the plurality of neighboring blocks are blocks spatially adjacent to the current block,
In the decoding, the plurality of neighboring blocks are decoded blocks,
In the decoding, the methods of generating a plurality of prediction blocks of the plurality of neighboring blocks are predictive block generation methods respectively used when generating a plurality of prediction blocks of the plurality of neighboring blocks,
In the decoding, a method of generating a prediction block for the current block is determined based on the selection information. The method of claim 18,
The selection information is a computer-readable recording medium indicating a method of generating a prediction block used to generate a prediction block of the current block. The method of claim 18,
The selection information is a computer-readable recording medium indicating a prediction block generation method used to generate a prediction block of the current block among the plurality of prediction block generation methods. The method of claim 18,
In the decoding, as the selection information indicates one prediction block generation method among the plurality of prediction block generation methods, a prediction block of the current block is generated by the one prediction block generation method. Recording medium. The method of claim 18,
In the decoding, a method of generating a prediction block of a neighboring block of the current block specified by the selection information is used as a method of generating a prediction block for the current block. The method of claim 18,
In the decoding, the prediction block generation method for the current block is determined based on the number of different values for the prediction block generation method used for predictions of the neighboring blocks. The method of claim 18,
In the decoding, the prediction for the current block is an intra prediction. The method of claim 18,
In the decoding, the prediction for the current block is a prediction using a specified decoded block as the prediction block for the current block.

Images