WO2011065727A2 - Moving image encoding/decoding apparatus and method through distortion block-based secondary prediction, and a recording medium therefor - Google Patents

Abstract

An embodiment of the present invention relates to a moving image encoding/decoding apparatus and method through distortion block-based secondary prediction, and a recording medium therefor, in which the prediction error is reduced by performing secondary prediction on the distortion blocks with reference to one of the distortion blocks. In particular, the method comprises: generating a primary prediction error signal based on an input signal and a predicted signal; performing adaptive DCT transform followed by first quantization on the primary prediction error signal; deciding whether secondary prediction is necessary, based on a distortion value of a signal that is obtained through inverse quantization followed by adaptive inverse DCT transform of the first quantization signal; generating at least one secondary prediction error signal for a corresponding block that needs the secondary prediction by referring to at least one representative block recorded in a lookup table, and then selecting one of the generated signals with a relatively low cost as an optimum, secondary prediction error signal; comparing a first cost of the primary prediction error signal with a second cost of the optimum, secondary prediction error signal; performing adaptive DCT transform followed by second quantization on the optimum, secondary prediction error signal, depending on the cost comparison result; and encoding the first quantized signal or the second quantized signal.

Description

Video encoding / decoding apparatus, method and recording medium using distortion block based second order prediction

An embodiment of the present invention relates to an image data compression technique. More specifically, a video through distortion block-based second prediction reduces a prediction error by performing second prediction on distortion blocks with reference to one of the distortion blocks. A coding / decoding apparatus, method and recording medium.

In general, the development of video compression technology has laid the foundation for a more efficient use of video media. In particular, H.264 / AVC video encoding technology has improved the compression performance by 2 times compared with the previous standard. This technique provides a coding step in the time and space domain based on the hybrid coding technique. The encoding step in the temporal domain reduces temporal redundancy through motion compensation prediction from an image of a reference frame. The motion compensation prediction is determined by the correlation between the block of the reference frame and the block of the image to be currently encoded, that is, the motion vector, and through this, a prediction image is obtained in units of blocks. The prediction error obtained as the difference between the predicted image and the original image is aligned in blocks, transformed into a frequency domain, and then quantized, and then scanned by a zigzag scanning method starting from a coefficient representing a DC value. Zigzag scanning produces an array of coefficients and subsequent coding steps can be optimized through CABAC or CAVLC. However, the coding efficiency is high by the DCT transform, which transforms the frequency domain only when the prediction errors in the block correlate with each other, that is, when they exist in the low frequency band. On the other hand, efficiency is low when only slightly correlated in the spatial domain.

Matthias Narroschke, Hans Georg Musmann "Adaptive prediction error coding in spatial and frequency domain for H.264 / AVC" VCEG-AB06, 16-20 January, 2006

In order to solve this problem, [Document 1] adds a method of encoding a prediction error in a spatial domain without performing a DCT transform to a method of encoding a prediction error in a conventional frequency domain, thereby making a prediction. A method of adaptively determining whether to convert an error signal into a frequency domain or to maintain a prediction error signal in a spatial domain for encoding is proposed.

1 is a flowchart illustrating a method of adaptively encoding a prediction error in [Document 1].

First, a prediction error signal of an image to be encoded is obtained through motion compensation prediction (S101).

For the prediction error obtained in step S101, DCT transform is performed, quantization is performed, and then quantization and DCT transform are inversely performed in the frequency domain based on the distortion and the required rate. Obtain the cost of (S102).

After the quantization is performed on the prediction error obtained in step S101, quantization is inversely performed to obtain a cost in the spatial domain based on distortion and a required rate (S103). ).

Finally, the cost in the frequency domain and the cost in the spatial domain, respectively, obtained in steps S102 and S103 are compared, and a coding method having a lower cost is selected to encode a prediction error signal (S104).

In the method of FIG. 1, it is assumed that when the DCT conversion is not performed, there is a more effective case than the DCT conversion.

The technique of encoding a prediction error signal by the method of FIG. 1 provided a higher encoding performance compared to the H.264 / AVC video encoding technique. However, if the prediction error sample in the block is not only low in the spatial domain but also large and small errors are scattered irregularly, the method is also less efficient.

An embodiment of the present invention is to solve the above-mentioned conventional problem, and its object is to perform second order prediction on the corresponding distortion blocks by referring to one of the distortion blocks in which the distortion in the first prediction is larger than the reference. Provided are a video encoding / decoding apparatus, method, and a recording medium through distortion block-based secondary prediction, which reduce prediction errors.

In order to achieve the above object, a video encoding apparatus using distortion block-based secondary prediction according to an aspect of the present invention includes a first prediction error signal generation unit generating a first prediction error signal based on an input signal and a prediction signal. A first quantizer configured to adaptively DCT transform the first prediction error signal and then quantize the first prediction error signal to a distortion value of a signal obtained by inverse quantization of the first quantized signal and then adaptively inverse DCT transform A second prediction determiner that determines whether the second prediction is necessary on a block basis based on one or more second prediction error signals with reference to each of one or more representative blocks recorded in the lookup table for the corresponding block requiring the second prediction; And select a relatively lowest cost signal as an optimal second prediction error signal, wherein the representative block requires the second prediction. A second prediction error signal generation unit comprising a block including at least a first block among all blocks; a cost comparison step of comparing a first cost of the first prediction error signal and a second cost of the corresponding optimal second prediction error signal A second quantizer for adaptively DCT-converting the optimal second prediction error signal and then performing second quantization according to the comparison result of the cost, and an encoder for encoding the first quantized signal or the second quantized signal. can do. If it is determined that the second cost is greater than the first cost by the cost comparison unit, the second prediction error signal generation unit replaces the block of the first prediction error signal corresponding to the first cost as a representative block in the lookup table. Can be added sequentially. The second prediction error signal generator may record, as reference block information, order information of the corresponding representative block referred to when generating the optimal second prediction error signal in the lookup table, and increase the reference frequency of the corresponding representative block. Then, it can be sorted based on the frequency.

In order to achieve the above object, the video decoding apparatus using the distortion block-based secondary prediction according to another aspect of the present invention provides a block discrimination unit for determining whether an input signal is a primary predictive encoded signal or a secondary predictive encoded signal in units of blocks. A first inverse quantizer for adaptively inverse DCT transformation after the first inverse quantization of the determined first prediction coded signal. If the determined first prediction coded signal is a representative block signal, the table storage unit sequentially stores the lookup table. A second inverse quantizer configured to adaptively inverse DCT transform the second inverse quantized signal after the second inverse quantized signal, and the lookup received according to the second prediction coded signal output from the second inverse quantizer A first-order prediction coded signal generator for generating a first-order prediction coded signal by referring to the corresponding representative block signal of the order information of the table; and And a decoder configured to decode the output signal of the first inverse quantizer or the output signal of the first predictive coded signal generator, wherein the table storage unit increases the reference frequency of the corresponding representative block signal referenced in the lookup table. It can be sorted based on the frequency.

In order to achieve the above object, a video encoding method using distortion block-based secondary prediction according to another aspect of the present invention may generate a first prediction error signal that generates a first prediction error signal based on an input signal and a prediction signal. A first quantization step of adaptively DCT transforming the first prediction error signal and then quantizing the first prediction error signal A distortion value and a predetermined reference value of a signal obtained by inverse quantization of the first quantized signal and then adaptively inverse DCT transforming A second prediction determining step of determining whether a second prediction is necessary on a block basis based on the at least one second prediction error for each corresponding block requiring the second prediction by referring to one or more representative blocks recorded in a lookup table, respectively; A signal is generated and a relatively lowest cost signal is selected as an optimal second prediction error signal, wherein the representative block is the second order. Generating a second prediction error signal comprising a block including at least a first block of all the blocks required for the side to compare the first cost of the first prediction error signal and the second cost of the corresponding second optimal prediction error signal Cost comparison step A second quantization step of adaptively DCT-converting and then quantizing the optimal second prediction error signal according to the comparison result of the cost and encoding the first quantized signal or the second quantized signal. An encoding step may be included. If it is determined in the cost comparison step that the second cost is greater than the first cost, further including a table updating step of sequentially adding a block of the first prediction error signal corresponding to the first cost to the lookup table as a representative block. can do. The updating of the table may include, as reference block information, order information of a corresponding representative block referred to when generating the optimal second prediction error signal in the lookup table, and after increasing the reference frequency of the corresponding representative block, the frequency You can sort by.

In order to achieve the above object, a video decoding method using distortion block-based secondary prediction according to another aspect of the present invention includes determining whether an input signal is a first prediction coded signal or a second prediction coded signal on a block basis. A first inverse quantization step of adaptively inverse DCT transformation after the first inverse quantization of the determined first prediction coded signal, and the table storing step of sequentially storing the determined first prediction coded signal in a lookup table if the determined first prediction coded signal is a representative block signal. A second inverse quantization step of adaptively inverse DCT transformation after the second inverse quantization of the determined second prediction coded signal; and the look-up received corresponding to the second prediction coded signal output from the second inverse quantization step Generate a first-order prediction coded signal generated as a first-order prediction coded signal by referring to the corresponding representative block signal in the order information of the table. In the lookup table, the reference frequency of the referenced representative block signal is increased, and then the table sorting step and the output signal of the first inverse quantization step or the output of the first predictive coded signal generation step are arranged. And a decoding step of decoding the signal.

According to still another aspect of the present invention, a computer-readable recording medium in which a video encoding method through the distortion block-based secondary prediction described above is recorded by a program may be provided.

According to still another aspect of the present invention, a computer-readable recording medium in which a video decoding method through the above-described distortion block-based secondary prediction is recorded by a program can be provided.

According to Document 1, the encoding error is adaptively encoded in the frequency domain and the spatial domain to provide higher encoding performance compared to the conventional H.264 / AVC video encoding technique. However, if the prediction error sample in the block is not only low in the spatial domain but also large and small errors are scattered irregularly, the method of [1] is still inefficient.

According to an embodiment of the present invention, since a block representing a group is selected from the first prediction error signals, additional information bits do not occur and the amount of computation can be reduced compared to the existing technology.

According to an embodiment of the present invention, even if adaptive encoding is performed in the existing frequency domain and the spatial domain, additional second prediction is performed for the case where the prediction error is large. To determine whether to perform additional secondary prediction, we distinguish between cases where the prediction error is large when performing the existing technique. The discrimination method according to an embodiment of the present invention is performed by comparing the distortion generated when the prediction error signal is encoded in the frequency and spatial domains with the threshold value K as a set reference value. If the distortion is smaller than the threshold value K, the prediction error is small and the compression efficiency is judged to be good. Therefore, coding is performed by the conventional method (APEC). If the distortion is larger than the threshold value K, the prediction error is determined to be larger. Make predictions.

According to an embodiment of the present invention, the second order prediction is performed by subtracting a prediction error signal of a reference block recorded in a lookup table for a block having a large distortion. The reference block is a block selected as a representative among previous blocks of the block to be currently encoded, and the cost RD_cost can be reduced by subtracting the prediction error signal of the representative (reference) block. The reference block can be used in the same way in further prediction for another block.

According to an embodiment of the present invention, the first prediction error signal having a distortion greater than the threshold value K in one slice or one frame is stored in the lookup table, which is set as the first reference block. The lookup table includes the frequency of use for the stored prediction error signal, and has the names of the blocks in the order of their order in the sorted order. Prediction error signals with large distortion after the first are subtracted from the prediction error signals of all reference blocks stored in the lookup table, and each cost is obtained through quantization and inverse quantization. Compare the previous costs. If the second prediction is not performed, that is, the cost after the first prediction is smaller, the prediction error signal of the block is added to the lookup table and used as a reference block. If the cost after the second prediction is smaller than the cost after the first prediction, the frequency of the prediction error signal of the lookup table used for the second prediction is increased, and the block is encoded using the prediction error signal of any block for decoding. The name is written on the prediction error signal to be encoded.

According to an embodiment of the present invention, an adaptive encoder is provided to use a complex encoding technique of a prediction error signal. According to an embodiment of the present invention, it is determined whether to encode in the frequency domain, the spatial domain, or the second-order prediction, and the determination is performed by a distortion weighted by a Lagrange parameter. The cost is calculated based on the cost and the rate, and the method of minimizing the cost is selected. According to an embodiment of the present invention, the request rate when performing additional secondary prediction includes a bit of an information block name to indicate a block used by the user in addition to the request rate calculation method of the existing technology.

According to an embodiment of the present invention, a side information signal is generated whether to be encoded in the frequency domain, the spatial domain, or after the additional prediction. For 8x8 blocks, each of the 4x4 blocks is divided into four 4x4 blocks. In order to indicate an encoding method, a total of 8 bits are required in 2 bits. An embodiment of the present invention uses pattern information on an encoding method to minimize such side information signals. Each of 4 4x4 blocks can be encoded using three methods (frequency domain, spatial domain, and additional prediction method), and thus, there are a total of 81 cases. Instead of the aforementioned 8 bits, 81 patterns can be used to represent a total of 7 bits. This minimizes side information signals to increase compression performance.

According to an embodiment of the present invention, a coding mechanism such as CABAC or the like is based on a probability determined separately for a coefficient in the frequency domain or a sample in the spatial domain.

According to an embodiment of the present invention, a decoder corresponding to the above-described encoding method is provided. It is determined whether to display a reference block of the encoded prediction error signal to determine whether to perform decoding according to the second-order prediction. If there is no reference block indication, it is decoded in the frequency or spatial domain in the same manner as in [Document 1]. If there is a reference block notation, the block name to be currently decoded is searched in the lookup table. If the block name does not exist in the lookup table, the block to be decoded is set as the reference block, and the prediction error signal is stored in the lookup table by the table storage unit. The prediction error signal is decoded in the frequency or spatial domain in the same manner as in [Document 1]. If there is a reference block notation and the block name already exists in the lookup table, the prediction error signal of the corresponding reference block and the prediction error signal of the block to be decoded are added to the inverse quantizer from the lookup table, and decoded by the decoding method selection signal. Sent to the flag.

According to various aspects of the present invention as described above, the distortion block-based 2 to reduce the prediction error by performing a second prediction for the distortion blocks with reference to one of the distortion blocks having a larger distortion than the first prediction By providing a video encoding / decoding method, apparatus, and a recording medium through differential prediction, if the prediction error is still large despite the adaptive coding in the existing frequency domain and the spatial domain, the distortion block based second prediction is performed. There is an effect of improving the compression efficiency by significantly reducing the prediction error.

1 is a flowchart illustrating a method of adaptively encoding a prediction error according to the prior art;

2 is a block diagram of a video encoding apparatus through distortion block-based second order prediction, according to an embodiment of the present invention;

3 is a detailed block diagram of a second prediction error signal generator of FIG. 2;

4 is a flowchart of a video encoding method using distortion block based second order prediction according to an embodiment of the present invention;

5 is a detailed flowchart of a second prediction error signal generation step of FIG. 4;

6 is a view showing an example of a lookup table according to an embodiment of the present invention;

7 is a diagram illustrating a second order prediction block according to an embodiment of the present invention;

8 is a block diagram of a video decoding apparatus through distortion block-based second order prediction, according to an embodiment of the present invention;

9 is a flowchart of a video decoding method using distortion block-based second order prediction 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. First of all, in adding reference numerals to the components of each drawing, it should be noted that the same reference numerals are used as much as possible even though they are shown in different drawings. In describing the present invention, when it is determined that the detailed description of the related well-known configuration or function may obscure the gist of the present invention, the detailed description thereof will be omitted.

In addition, in describing the component of this invention, terms, such as 1st, 2nd, A, B, (a), (b), can be used. These terms are only for distinguishing the components from other components, and the nature, order or order of the components are not limited by the terms. If a component is described as being "connected", "coupled" or "connected" to another component, that component may be directly connected to or connected to that other component, but there may be another configuration between each component. It is to be understood that the elements may be "connected", "coupled" or "connected".

FIG. 2 is a block diagram of a video encoding apparatus through distortion block based second order prediction, according to an embodiment of the present invention. As shown in the drawing, a first prediction error signal generator 210 and a first quantizer ( 220, inverse quantizer 230, second prediction determiner 240, second prediction error signal generator 250, cost comparer 260, second quantizer 270, adaptive controller 280 ), And an entropy encoder 290.

The first prediction error signal generator 210 subtracts the prediction signal from the input signal to generate a first prediction error signal.

The first quantizer 220 adaptively performs DCT conversion on the first prediction error signal and then quantizes the first quantization unit. The DCT converter 221 performs DCT conversion on the first prediction error signal, and quantizes the DCT converted signal. A quantization unit 222 and a quantization unit 223 for quantizing the first prediction error signal without DCT conversion.

The inverse quantizer 230 inversely quantizes the output signal of the first quantizer 220 and then inversely DCT transforms the inverse quantizer 231 for inversely quantizing the output signal of the quantizer 222. An inverse DCT converter 232 for inverse DCT transforming the inverse quantized signal, and an inverse quantizer 233 for inverse quantizing the output signal of the quantization unit 223.

The second prediction determiner 240 determines whether the second prediction is necessary in units of blocks based on the distortion value of the output signal of the inverse quantizer 230 and a preset reference value. As shown in FIG. The input signal reconstruction unit 241 for reconstructing the input signal by adding the prediction signal to the output signal of the inverse quantization unit 230, compares the original input signal with the reconstructed input signal, obtains a distortion value as a difference therebetween, and the distortion The block of the first prediction signal whose value is greater than or equal to the reference value is determined to be a corresponding block requiring second prediction.

The second prediction error signal generator 250 generates one or more second prediction error signals with respect to the corresponding block requiring second prediction by referring to one or more representative blocks recorded in the lookup table, and among them, the lowest cost. Is selected as an optimal second prediction error signal, wherein the representative block may be composed of at least one or more blocks among all blocks requiring second prediction. 6 illustrates an example of a lookup table according to the present embodiment, and may include a block name as a block order of a 4x4 representative block, a block as a prediction error, and a reference frequency.

The cost comparison unit 260 compares the first cost of the first prediction error signal and the second cost of the corresponding optimal second prediction error signal, and the second prediction error signal generator 250 compares the cost comparison unit 260. If it is determined that the second cost obtained in the) is greater than the first cost, blocks of the first prediction error signal corresponding to the first cost may be sequentially added to the lookup table as a representative block. In addition, the second prediction error signal generator 250 records, as reference block information, order information of the corresponding representative block referred to when generating the optimal second prediction error signal in the lookup table, and increases the reference frequency of the corresponding representative block. After that, you can sort by that frequency.

When the second quantizer 270 determines that the second cost in the second prediction is smaller than the first cost in the first prediction, as a result of the cost comparison in the cost comparison unit 260, the second quantization unit 270 generates the second prediction error signal accordingly. The second-order quantization is performed by adaptively DCT transforming the optimal second prediction error signal output from 250.

 The adaptive control unit 280 quantizes the first prediction error signal in the frequency domain, the signal A quantizes the first prediction error signal in the spatial domain, and the optimal second prediction error signal in the spatial domain. One of the signal C and the signal D obtained by quantizing the optimal second prediction error signal in the frequency domain can be selected based on the distortion and the required ratio.

The entropy encoder 290 entropy encodes one signal selected by the adaptive controller 280 among the signals A, B, C, and D.

3 is a detailed block diagram of the second prediction error signal generator 250 of FIG. 2. The table search unit 251, the second prediction error signal generator 253, and the optimal second prediction error signal selector 255 are illustrated in FIG. 2. And a table update and additional information generator 257.

The table search unit 251 sequentially searches and outputs one or more representative blocks stored in the lookup table.

The second prediction error signal generation unit 253 includes a current block (or called a distortion block) and a table search unit 251 of the first prediction error signal classified by the distortion comparison unit 242 as requiring the second prediction. A block of one or more second prediction error signals is generated by obtaining a difference from one or more representative blocks sequentially output from the N-th prediction block.

The optimal second prediction error signal selector 255 optimizes one block having a relatively minimum cost among the blocks of one or more second prediction error signals generated for the current distortion block from the second prediction error signal generator 253. Selected as the second prediction error signal.

The table update and additional information generator 257 determines that the second prediction cost for the optimal second prediction error signal is greater than the corresponding first prediction cost based on the cost obtained by the cost comparison unit 260. A block of the prediction error signal is sequentially added to the lookup table as a representative block, the order information of the representative block referred to when generating the optimal second prediction error signal in the lookup table is recorded as the reference block information, and the reference of the representative block is referred to. After increasing the frequency, sort by that frequency.

FIG. 4 is a flowchart of a video encoding method through distortion block-based second order prediction according to an embodiment of the present invention. As an example, the present invention may be applied to the apparatus of FIG. 2 and will be described in parallel with the operation of the apparatus.

In the method of FIG. 4, as illustrated in the drawing, the first prediction error signal generation step S410, the first quantization step S420, the second prediction determination step S430, and the second prediction error signal generation step ( S440), a cost comparison step S450, a second quantization step S460, and an encoding step S470.

The first prediction error signal generation step S410 generates a first prediction error signal based on the input signal and the prediction signal through the first prediction error signal generation unit 210, and the first quantization step S420 is performed by the first prediction error signal generation unit 210. The first quantization unit 220 first adaptively transforms the first prediction error signal and then quantizes the first prediction error signal.

In the second prediction determination step S430, after inverse quantization of the first quantized signal through the inverse quantization unit 230, inverse DCT transformation is adaptively performed, and the inverse quantization unit 230 is performed through the second prediction determination unit S430. On the basis of the distortion value of the output signal and the predetermined reference value to determine whether or not secondary prediction is required in units of blocks, the prediction signal is added to the output signal of the inverse quantizer 230 through the input signal reconstruction unit 241 The input signal is reconstructed, the distortion comparison unit 242 compares the original input signal with the reconstructed input signal, obtains a distortion value as a difference between them, and then removes the block of the first prediction signal whose distortion value is equal to or greater than the reference value. It is determined that the block that requires the difference prediction.

The second prediction error signal generation step S440 may be performed by referring to one or more representative blocks recorded in the lookup table for the corresponding block requiring second prediction through the second prediction error signal generation unit 250, respectively. Generates an error signal and selects the least expensive signal as the optimal second prediction error signal, and also uses the order information of the corresponding representative block referred to when generating the optimal second prediction error signal in the lookup table as reference block information. The reference frequency of the corresponding representative block is increased, and then sorted based on the frequency.

The cost comparison step S450 compares the cost of the first cost of the first prediction error signal with the second cost of the corresponding optimal second prediction error signal through the cost comparison unit 260, and the cost comparison unit 260. If it is determined that the second cost obtained in E is greater than the first cost, the second prediction error signal generator 250 sequentially adds the blocks of the first prediction error signal corresponding to the first cost to the lookup table as representative blocks.

In the second quantization step S460, when it is determined that the second cost in the second prediction is smaller than the first cost in the first prediction, as a result of the cost comparison in the cost comparison unit 260, the second prediction error signal generator 250 is used. DCT transforms the optimal second prediction error signal output from the second through the second quantizer 270 and then quantizes it.

In the encoding step (S470), the adaptive control unit 280 may include a signal A quantized the first prediction error signal in the frequency domain, a signal B quantized the first prediction error signal in the spatial domain, and an optimal second prediction error. If one of the signal C quantized in the spatial domain and the signal D quantized the optimal second prediction error signal in the frequency domain are selected based on the distortion and the required ratio, the entropy encoder 290 may be used. Entropy-encodes one selected signal among the signals A, B, C, and D.

FIG. 5 is a detailed flowchart of the second prediction error signal generation step S440 of FIG. 4, which may be applied to the apparatus of FIG. 3 as an example and will be described in parallel with the operation of the apparatus.

5 is a table search step (S510), a second prediction error signal generation step (S520), an optimal second prediction error signal selection step (S530), a cost comparison step (S540), a table update and additional information generation step ( S550).

Table Search Steps (S510)

First, the table search unit 251 sequentially searches for a prediction error signal (or a representative block) stored in the lookup table (S511), and determines whether a prediction error signal stored in the lookup table exists (S512). As a result, when it is determined that the prediction error signal exists in the lookup table, it is determined whether a prediction error signal which is not used for the second prediction among the prediction error signals exists (S513).

Second prediction error signal generation step (S520)

Subsequently, when it is determined in step S513 that a prediction error signal that is not used for secondary prediction exists, the secondary prediction error signal generator 253 may predict the corresponding prediction error signal of the lookup table from the primary prediction error signal to be currently encoded. Is subtracted to generate a second prediction error signal. Steps S510 to S511 are performed until all prediction error signals in the lookup table are used as reference for the second prediction with respect to the first prediction error signal to be currently encoded.

Optimal Second Prediction Error Signal Selection Step (S530)

Subsequently, the optimal second prediction error signal selector 255 quantizes and dequantizes the second prediction error signal generated by the second prediction error signal generator 253 in step S520 (S531), and calculates a cost ( S532), after determining whether the previously calculated cost is greater than the current calculated cost (S533), if the previously calculated cost is large, the previous calculation cost and the second prediction error signal are deleted, and the current calculation cost and the secondary The prediction error signal is stored (S534).

The above-described processes of steps S510-S530 are repeatedly performed on the first prediction error signal to be encoded until all prediction error signals in the lookup table are used as reference for the second prediction, thereby generating at least one second prediction. The optimal second prediction error signal having the lowest cost among the error signals is selected by the optimum second prediction error signal selection unit 255.

Cost comparison step (S540)

The cost comparison step S540 is a cost if the prediction error signal not used for the second prediction in step S513 does not exist, that is, if all prediction error signals in the lookup table are used as reference for the second prediction for the current block. The comparison unit 260 compares the first cost of the first prediction error signal with the second cost of the corresponding optimal second prediction error signal (ie, the last stored cost in step S534).

Table update and additional information generation step (S550)

If it is determined in step S512 that the prediction error signal stored in the table does not exist, or in step S540 it is determined that the first cost of the primary prediction error signal is less than the second cost of the corresponding optimal secondary prediction error signal, the second prediction error. The block of the first prediction error signal corresponding to the first cost is sequentially added to the lookup table as the representative block through the table ganging and the additional information generator 257 of the signal generator 250 (S551), and the added representative The sequence number of the block is given as the block name (S552). In addition, if it is determined in step S540 that the first cost of the difference prediction error signal is greater than the second cost of the corresponding optimal second prediction error signal, the table ganging and additional information generation unit of the second prediction error signal generator 250 ( 257) records, as reference block information, the order information of the representative block referenced when generating the optimal second prediction error signal in the lookup table (S553), and then increases the reference frequency of the corresponding representative block and sorts based on the frequency. (S554).

Subsequently, an apparatus and method for encoding a video through vector quantization based secondary prediction according to an embodiment of the present invention will be described with reference to FIGS. 2 to 7.

Embodiments of the present invention are based on H.264 / AVC technology, but may be performed in combination with existing technology.

According to an exemplary embodiment of the present invention, the motion compensation prediction is performed based on the motion estimation to provide the prediction signal after the input signal undergoes the motion estimation, and the first prediction signal generator 210 receives the prediction signal from the input signal. Subtract to generate the first prediction error signal. The first prediction error signal generated therefrom is converted into a frequency domain by the DCT unit 221 and quantized by the quantization unit 222. The output signal of the quantization unit 222 is inverse quantized and inverse DCT changed through the inverse quantization unit 231 and the inverse DCT unit 232 and used for motion compensation prediction. In addition, the first prediction error signal may be quantized by being sent to the quantization unit 223 in the spatial domain without being transformed into the frequency domain, and then adaptively select the first prediction error signals A and B encoded in the frequency domain and the spatial domain. Make sure In addition, the first prediction error signal is added to the prediction signal through the input signal reconstruction unit 241 and used for motion compensation and motion estimation.

According to an embodiment of the present invention, additional second prediction is performed only when the distortion of the quantized first prediction error signal and the quantized first prediction error signal in the spatial domain is large. Determination of whether to perform the second prediction is performed by the distortion comparator 242, and is determined by comparing the quantized primary prediction error signal and the threshold value K in the frequency domain and the spatial domain which are inputs of the distortion comparator 242. The threshold K is a criterion for the large and small distortion. When the distortion of the quantized first prediction error signal is smaller than the threshold value K, the encoding is performed in the frequency domain or the spatial domain in the same manner as in the conventional technique, under the determination that the distortion is small and the compression efficiency is good. On the other hand, if the distortion of the quantized first prediction error signal is greater than K, the second prediction is performed under the determination that the distortion is large and the compression efficiency is lowered to increase the compression efficiency. For all blocks such as A of FIG. 7, a block having a large distortion of the first prediction error signal is classified by determining whether a second prediction is necessary, such as B, through the above-described comparison method. The hatched block of B denotes a block for performing additional prediction due to large distortion, and the first prediction error signal is sent to the second prediction error signal generator 250 for the second prediction.

3 and 5, when the first prediction error signal having a large distortion enters the second prediction error signal generator 250, the table searcher 251 searches for the prediction error signal stored in the table. . If the prediction error signal to be currently encoded is to perform second prediction for the first time, no data exists in the lookup table. Because secondary prediction cannot be performed, the block of the first primary prediction error signal to be encoded currently is stored directly in the lookup table as a representative block, which is later used for secondary prediction of other prediction error signals. It is given as', and denotes a representative block as shown in C of FIG. 7 to indicate that the block used for the second prediction for the first prediction error signal. If the prediction error signal to be encoded currently does not perform additional prediction for the first time, one or more prediction error signals are stored in the lookup table. The second prediction error signal generator 253 generates a second prediction error signal by obtaining a difference between the prediction error signal stored in the lookup table and the first prediction error signal of the block to be currently encoded (S520). After the quantization and inverse quantization are performed and the prediction error signal of the representative block stored in the lookup table is added again (S531), the cost of performing the second prediction is calculated (S532). By comparing the cost (S533), if the calculated cost is greater than the cost through the first prediction, the second prediction result is ignored. On the contrary, if the cost through the first prediction is smaller, the cost through the first prediction is the second prediction. It is replaced with the result of performing the operation (S534). When the process is completed, the second prediction is performed by referring to the next representative block stored in the table search unit 251 again. The above process is performed by referring to all the representative blocks stored in the lookup table for the block to be encoded, and the cost of the lowest cost result is stored and the corresponding second prediction error signal is the final second order. The prediction error signal is selected (S534). In addition, the secondary prediction error signal of the representative block used for the final secondary prediction error signal is set as a reference prediction error signal block (or referred to as a reference block), and the corresponding reference block name corresponds to the primary prediction error signal to be currently encoded. In addition, it is encoded. If the second prediction is performed even though the prediction error signal of the representative block stored in the lookup table is higher than the cost of performing the first prediction, the result of performing the second prediction is not used, and the prediction error to be currently encoded The signal is added to the lookup table as a representative block and this representative block information is indicated.

6 is a diagram illustrating an example of a lookup table according to an exemplary embodiment of the present invention. The table not only stores prediction error signals, but also stores names and frequencies of blocks. Each time it is used for secondary prediction, the frequency increases, and the table is sorted in order based on this frequency. The block names are assigned in order from 0 in this order, and as the order increases, the number of bits for indicating the block names increases. The most frequent prediction error signal located at the top of the lookup table has the block name '0' and 1 bit is required to indicate this information. For example, the sixth prediction error signal has the block name '101' and 3 I need a bit.

According to an embodiment of the present invention, the second prediction is performed only when the distortion of the quantized prediction error signal in the frequency domain and the quantized prediction error signal in the spatial domain is large. If the cost of performing the second prediction after performing the second prediction is greater than the cost of the first prediction, the corresponding first prediction error signal is stored in the lookup table, and by the adaptive controller 280, according to the first prediction One of A and B is selected as the encoding method in the frequency domain or the spatial domain. If the cost according to the second prediction is smaller than the cost according to the first prediction, the encoding method C or D for performing the second prediction by the adaptive controller 280 is selected. The cost by the second prediction includes the cost for representing the block name described above.

In FIG. 2, A denotes a first prediction error signal quantized in the frequency domain, B denotes a first prediction error signal quantized in the spatial domain, C denotes an optimal second prediction error signal quantized in the spatial domain, and D denotes a frequency prediction. It is a quantized optimal second order prediction error signal. The adaptive controller 280 selects one prediction error signal that minimizes the cost among the prediction error signals A, B, C, and D encoded by four methods, and the entropy encoder 280 selects one selected signal. Encode

FIG. 8 is a block diagram of a video decoding apparatus through distortion block based second order prediction, corresponding to the encoding apparatus of FIG. 2, and including a block discriminator 810 and a first inverse quantizer 820. , A table storage unit 830, a first-order prediction coded signal generator 840, a second inverse quantizer 850, and a decoder 860.

The block discriminating unit 810 determines whether the input signal is a primary predictive encoded signal or a secondary predictive encoded signal based on additional information (whether the representative block and the reference block are written) on a block basis.

The first inverse quantizer 820 adaptively inverses DCT transforms the determined first predictive coded signal after the first inverse quantization. If the corresponding signal is a quantized coded signal in the frequency domain, the inverse quantizer ( Inverse quantization through 821 and outputting signal A 'after inverse DCT conversion through inverse DCT conversion unit 822, and through inverse quantization unit 823 to correspond to the corresponding signal if it is a coded signal quantized in a spatial domain. The signal B 'after inverse quantization is output.

The table storage unit 830 sequentially stores the determined first prediction coded signal in the lookup table when the determined first prediction coded signal is a representative block signal. In addition, the table storage unit 830 increases the reference frequency of the corresponding representative block signal referenced in the lookup table and then sorts the frequency based on the frequency. The configuration of the lookup table of the present embodiment is the same as that of the lookup table used in the above-described encoding.

The first predictive coded signal generator 840 may determine the determined second predictive coded signal (or an output signal of the second inverse quantizer 850), and corresponding representative block signal of the order information of the received lookup table corresponding thereto. By reference, a first prediction coded signal is generated.

 The second inverse quantizer 850 adaptively inverses the second inverse quantized signal generated by the first predictive coded signal generator 840 and outputs the second predicted coded signal (or the determined second predictive coded signal). DCT conversion, and if the corresponding signal is a quantized coded signal in the spatial domain, the signal C 'after inverse quantization is output through the inverse quantization unit 851, and if the signal is a quantized coded signal in the frequency domain, Correspondingly, the inverse quantization is performed through the inverse quantization unit 852 and the inverse DCT conversion is performed through the inverse DCT conversion unit 853.

The decoder 860 decodes one signal selected by the decoding method selection signal among the signals A ', B', C ', and D'. The decoding method selection signal may be generated through cost comparison with respect to the signals A ', B', C ', and D' corresponding to the adaptive control unit 280 of FIG.

Meanwhile, according to another embodiment of the present invention, unlike the FIG. 8, the second inverse quantizer 850 is positioned at the tip of the first predictive coded signal generator 840 to inversely quantize and adaptively inverse the input signal. After DCT conversion, the output signal can be used as an input of the first-order predictive coded signal generator 840.

FIG. 9 is a flowchart of a video decoding method through distortion block-based second order prediction according to an embodiment of the present invention. As an example, the present invention may be applied to the apparatus of FIG. 8 and will be described in parallel with the operation of the apparatus.

First, the block determining unit 810 determines whether the input signal is a first prediction coded signal or a second prediction coded signal on a block-by-block basis (S910), and the determined first prediction coded signal is converted into a first inverse quantizer ( The first inverse quantization is performed through 820 to adaptively perform inverse DCT conversion to output signals A 'and B' (S920).

Subsequently, if the primary prediction coded signal determined by the block determining unit 810 is the representative block signal, the representative block signal is sequentially stored in the lookup table through the table storage unit 830 (S930).

Next, the first prediction error signal generator 840 may output the second prediction coded signal ( or the second inverse quantizer 850) determined by the block discriminator 810 with reference to the representative block signal of the lookup table. A signal is generated as a first-order prediction coded signal, and the representative block signal of the order information of the lookup table as the received additional information is added to the second-order prediction coded signal to generate a first-prediction coded signal (S940).

Next, the table storage unit 830 increases the reference frequency of the representative block signal referred to in the lookup table and then sorts the frequency based on the frequency (S950).

Next, the second inverse quantizer 850 adaptively performs the second inverse quantization of the first prediction coded signal (or the determined second prediction coded signal) generated and output from the first prediction coded signal generator 840. Inverse DCT conversion, for example, if the signal is a quantized coded signal in the spatial domain, the signal C 'after inverse quantization is output through the inverse quantizer 851 so as to correspond thereto, and the signal is quantized in the frequency domain. In response to this, inverse quantization is performed through the inverse quantization unit 852, and the signal D ′ after the inverse DCT conversion is output through the inverse DCT conversion unit 853 (S960).

Finally, the decoding unit 860 decodes one signal selected by the decoding method selection signal among the signals A ', B', C ', and D', and the decoding method selection signal is the adaptive control unit of FIG. Corresponding to 280, the signal is generated through the cost comparison for the signals A ', B', C ', and D' (S1005).

Meanwhile, according to another embodiment of the present invention, unlike step 9, step S960 may be performed before step S940 and step S940 may be performed.

Subsequently, an operation of a video decoding apparatus through distortion block based second order prediction and a method corresponding thereto will be described with reference to FIGS. 8 and 9.

According to an embodiment of the present invention, a decoding apparatus / method corresponding to the above-described encoding apparatus / method is provided. The decoding apparatus of FIG. 8 and the decoding method of FIG. 9 according to the present embodiment provide the same lookup table as the lookup table provided by the encoder of FIG. 2 for decoding. The coded prediction error signal determines whether the second prediction is performed by the block marking determination unit 810. If there is no reference block notation, the first inverse quantizer 820 of FIG. 8 is selected and decoded in the frequency domain or the spatial domain as shown in [Document 1]. When the reference block representation represents a representative prediction error signal, that is, when the prediction error signal of the block to be decoded is used as the representative block, the corresponding prediction error signal is stored in the lookup table by the table storage unit 830. The first inverse quantization unit 820 is selected and decoded in the same manner as in [Document 1]. When there is a reference block notation that the block to be decoded refers to the block stored in the lookup table for the second prediction, the first prediction coded signal generator 840 of FIG. 8 is selected and the first prediction coded signal. The generator 840 generates a first predictive coded signal by adding a prediction error signal of a corresponding reference block and a prediction error signal of a block to be decoded from the lookup table of the table storage unit 830, and then generates a second inverse quantizer. The signal is sent to the decoding unit 860 by the decoding method selection signal.

According to an embodiment of the present invention, decoding and scanning according to CABAC or CAVLC are performed by the same method as the existing technology.

The video encoding method through the distortion block based second order prediction according to the embodiment of the present invention described with reference to FIGS. 4-5 and the distortion block based second order prediction according to the embodiment described with reference to FIG. 9. The video decoding method may be implemented as a computer readable recording medium including program instructions for performing operations implemented by various computers. The computer readable recording medium may include program instructions, local data files, local data structures, etc. alone or in combination. The recording medium may be those specially designed and constructed for the embodiments of the present invention, or may be known and available to those skilled in computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tape, optical recording media such as CD-ROMs, DVDs, magnetic-optical media such as floppy disks, and ROM, RAM, flash memory, and the like. Hardware devices specifically configured to store and execute the same program instructions are included. The recording medium may be a transmission medium such as an optical or metal wire, a waveguide, or the like including a carrier wave for transmitting a signal specifying a program command, a local data structure, or the like. Examples of program instructions may include high-level language code that can be executed by a computer using an interpreter as well as machine code such as produced by a compiler.

In the above description, all elements constituting the embodiments of the present invention are described as being combined or operating in combination, but the present invention is not necessarily limited to the embodiments. In other words, within the scope of the present invention, all of the components may be selectively operated in combination with one or more. In addition, although all of the components may be implemented in one independent hardware, each or all of the components may be selectively combined to perform some or all functions combined in one or a plurality of hardware. It may be implemented as a computer program having a. Codes and code segments constituting the computer program may be easily inferred by those skilled in the art. Such a computer program may be stored in a computer readable storage medium and read and executed by a computer, thereby implementing embodiments of the present invention. The storage medium of the computer program may include a magnetic recording medium, an optical recording medium, a carrier wave medium, and the like.

In addition, the terms "comprise", "comprise" or "having" described above mean that the corresponding component may be included, unless otherwise stated, and thus excludes other components. It should be construed that it may further include other components instead. All terms, including technical and scientific terms, have the same meanings as commonly understood by one of ordinary skill in the art unless otherwise defined. Terms commonly used, such as terms defined in a dictionary, should be interpreted to coincide with the contextual meaning of the related art, and shall not be construed in an ideal or excessively formal sense unless explicitly defined in the present invention.

The above description is merely illustrative of the technical idea of the present invention, and those skilled in the art may make various modifications and changes without departing from the essential characteristics of the present invention. Therefore, the embodiments disclosed in the present invention are not intended to limit the technical spirit of the present invention but to describe the present invention, and the scope of the technical idea of the present invention is not limited by these embodiments. The scope of protection of the present invention should be interpreted by the following claims, and all technical ideas within the scope equivalent thereto should be construed as being included in the scope of the present invention.

As described above, an embodiment of the present invention is applied to a field of image data compression technology, and when distortion of a first prediction error signal is large, distortion block-based second prediction is performed and DCT transformation is adaptively performed to reduce prediction error. It is a very useful invention that can be significantly reduced.

CROSS-REFERENCE TO RELATED APPLICATION

This patent application claims priority under patent application number 119 (a) (35 USC § 119 (a)) to patent application No. 10-2009-0113919, filed with Korea on 24 November 2009. All content is incorporated by reference in this patent application. In addition, if this patent application claims priority for the same reason as above for a country other than the United States, all the contents thereof are incorporated into this patent application by reference.

Claims (10)

1. A first prediction error signal generator configured to generate a first prediction error signal based on the input signal and the prediction signal;

   A first quantizer for adaptively DCT transforming the first prediction error signal and then performing first quantization;

   A second prediction determination unit determining whether a second prediction is necessary in units of blocks based on a distortion value of a signal obtained by inverse quantization of the first quantized signal and adaptively performing inverse DCT conversion and a predetermined reference value;

   For the corresponding block requiring the second prediction, one or more second prediction error signals are generated by referring to each of the one or more representative blocks recorded in the lookup table, and among them, the lowest cost signal is converted into the optimal second prediction error signal. A second prediction error signal generator comprising a block including at least a first block among all blocks requiring second prediction;

   A cost comparison unit for comparing a first cost of the first prediction error signal with a second cost of the corresponding optimal second prediction error signal;

   A second quantizer for adaptively DCT-converting the optimal second prediction error signal according to the comparison result of the cost and then performing second quantization; And

   An encoder which encodes the first quantized signal or the second quantized signal;

   An apparatus for encoding a video using distortion block-based secondary prediction, comprising a.
2. The method of claim 1,

   If it is determined that the second cost is greater than the first cost by the cost comparison unit, the second prediction error signal generator is a block of the first prediction error signal corresponding to the first cost as a representative block in the lookup table. A video encoding apparatus using distortion block-based second order prediction, which is added sequentially.
3. The method of claim 2,

   The second prediction error signal generator records, as reference block information, order information of the corresponding representative block referred to when generating the optimal second prediction error signal in the lookup table, and increases the reference frequency of the corresponding representative block. And a video encoding apparatus based on distortion block based second order prediction, wherein the frequency is aligned based on the frequency.
4.  A block discrimination unit determining whether the input signal is a primary predictive encoded signal or a secondary predictive encoded signal in units of blocks;

   A first inverse quantizer for adaptively inverse DCT transforming the first predictive coded signal after first inverse quantization;

   A table storage unit sequentially storing the determined first prediction coded signal in a lookup table if the determined first prediction coded signal is a representative block signal;

   A second inverse quantizer for adaptively inverse DCT transforming the second predicted coded signal after inverse second quantization;

   Generating a first-order prediction coded signal to generate a second-order prediction coded signal output from the second inverse quantization unit as a first-order prediction coded signal by referring to a corresponding representative block signal of the order information of the lookup table received correspondingly. part; And

   A decoder which decodes an output signal of the first inverse quantizer or an output signal of the first predictive coded signal generator;

   Including,

   And the table storage unit increases the reference frequency of the corresponding representative block signal referred to in the lookup table and then sorts the frequency based on the frequency.
5. Generating a first prediction error signal based on the input signal and the prediction signal;

   A first quantization step of adaptively DCT transforming the first prediction error signal and then performing first quantization;

   A second prediction determining step of determining whether a second prediction is required in units of blocks based on a distortion value of a signal obtained by inverse quantization of the first quantized signal and then adaptively performing inverse DCT conversion and a predetermined reference value;

   For the corresponding block requiring the second prediction, one or more second prediction error signals are generated by referring to each of the one or more representative blocks recorded in the lookup table, and among them, the lowest cost signal is converted into the optimal second prediction error signal. Selecting, wherein the representative block comprises a second prediction error signal generation step comprising a block including at least a first block among all blocks requiring second prediction;

   A cost comparison step of comparing a first cost of the first prediction error signal and a second cost of the corresponding second best prediction error signal;

   A second quantization step of adaptively DCT-converting the second optimal prediction error signal according to the comparison result of the cost and then performing second quantization; And

   An encoding step of encoding the first quantized signal or the second quantized signal;

   Video encoding method through the distortion block-based secondary prediction, comprising a.
6. The method of claim 5,

   If it is determined in the cost comparison step that the second cost is greater than the first cost, further including a table updating step of sequentially adding a block of the first prediction error signal corresponding to the first cost to the lookup table as a representative block. A video encoding method based on distortion block based second order prediction.
7. The method of claim 6,

   The updating of the table may include, as reference block information, order information of a corresponding representative block referred to when generating the optimal second prediction error signal in the lookup table, and after increasing the reference frequency of the corresponding representative block, the frequency The video encoding method through the distortion block-based second-order prediction, characterized in that aligned on the basis of.
8. A discriminating step of determining whether the input signal is a first prediction coded signal or a second prediction coded signal in units of blocks;

   A first inverse quantization step of adaptively inverse DCT transforming the first predictive coded signal after first inverse quantization;

   A table storing step of sequentially storing the determined first prediction coded signal in a lookup table if the determined first prediction coded signal is a representative block signal;

   A second inverse quantization step of adaptively inverse DCT transformation after second inverse quantization of the determined second prediction coded signal;

   Generating a first-order prediction coded signal to generate a second-order prediction coded signal output from the second inverse quantization step by referring to a corresponding representative block signal of order information of the lookup table received correspondingly. step;

   A table sorting step of increasing a reference frequency of the referenced representative block signal in the lookup table and then sorting the frequency based on the frequency; And

   A decoding step of decoding the output signal of the first inverse quantization step or the output signal of the first predictive coded signal generation step;

   Video decoding method through the distortion block-based secondary prediction, comprising a.
9. A computer-readable recording medium in which a video encoding method using the distortion block based second order prediction according to any one of claims 5 to 7 is recorded by a program.
10. The computer-readable recording medium of claim 8, wherein the video decoding method through the distortion block-based second order prediction is recorded by a program.

Images