KR20110022933A - Deblocking filtering unit with a reconfigurable data-path and method thereof - Google Patents

Abstract

PURPOSE: A deblocking filter calculation device and a method thereof including a common data pass structure are provided to integrate a data pass which performs strong intensity filtering and a data pass performing average intensity filtering. CONSTITUTION: A critical value generating unit generates a critical value from an inputted index value. An edge judging unit determines whether an edge, which is a current filtering target, is made due to an error because of a blocking artifact or not by using an input pixel data which will be filtered with the generated critical value. A filtering calculation performing unit selectively performs an average intensity filtering calculation or a strong strength filtering calculation according to an inputted boundary intensity coefficient. An outputting unit outputs an average intensity filtering calculation result or a strong intensity filtering calculation result.

Description

Deblocking filtering unit with a reconfigurable data-path and method thereof

The present invention relates to an apparatus and method for deblocking filtering operation including a common data path structure, and more particularly, to a deblocking filtering operation having a configuration in which a data path for performing normal intensity filtering and a data path for performing strong intensity filtering are integrated. An apparatus and a method thereof are provided.

Standardized by MPEG-2 and MPEG-4, video compression codecs standardized by the Motion Picture Experts Group (MPEG) under the World Organization for Standardization (ISO), and by the Video Coding Experts Group (VCEG) under the International Telecommunication Union (ITU-T). A video compression codec such as H.263 and H.264 performs encoding and decoding processes based on a block having a predetermined size. In the block-based video compression codec, a blocking artifact phenomenon occurs that degrades the subjective and objective image quality of the image, and a deblocking filter is used in an in-loop or post manner to remove the blocking artifact.

In the H.264 / AVC codec, a block having a size of 16 × 16 is used as a basic unit of image encoding and decoding, which is called a macroblock. A video codec encoding process in units of macroblocks will be described.

FIG. 1 is a block diagram of a video codec encoder in a macroblock unit.

The video codec encoder 100 may include a discrete cosine transform (DCT) unit 110, a quantization unit 120, a motion predictor 130, a motion compensator 140, an entropy encoding and decoding unit 150, and an inverse quantization unit. 160, an inverse DCT unit 170, and a deblocking filter 180.

First, the video codec encoder 100 searches for a portion of the reference video frame having the smallest difference from the macroblock through intra or inter prediction on the macroblock of the original video frame to be encoded. After that, the residual with the most similar part is obtained. The DCT unit 110 performs a discrete cosine transform by inputting the difference value, and the quantization unit 120 performs a discrete cosine transform on the result. Perform the quantization process.

The entropy encoding and decoding unit 150 performs entropy encoding on the result of the quantization process and finally generates a compressed H.264 / AVC bitstream.

On the other hand, the video codec encoder 100 performs the reconstruction process in the same procedure as the decoding process to generate a reference image frame which will be needed in the next compression process.

That is, the entropy encoding and decoding unit 150 performs an entropy decoding process on the compressed H.264 / AVC bitstream, and the dequantization unit 160 dequantizes it. On the result, the inverse DCT unit 170 performs inverse discrete cosine transform to restore residual information. The difference value information reconstructed as described above is decoded by being combined with a macro block subjected to intra or inter motion compensation.

The decoded image frame expresses a lattice boundary error that does not exist in the original image at the macroblock or subdivided macroblock boundary. The error is a phenomenon in which abnormal cosine transforms are relatively inaccurate for block boundary pixels due to discrete cosine transforms in blocks, lossy compression is performed through quantization, and restored to a value different from the original image in inverse quantization. This is caused by a phenomenon that occurs due to block prediction.

Such blocking artifacts recommend the use of the deblocking filter 180 in modern video compression codecs in order to improve reconstruction quality due to subjective picture quality degradation. In particular, in H.264 / AVC, the deblocking filter is included as a standard so that the deblocking filter is performed in a video encoding and decoding loop.

The deblocking filter 180 of H.264 / AVC distinguishes the edge existing in the original image from the edge generated by the blocking artifact, and does not deblock the edge existing in the original image. have.

The H.264 / AVC video compression codec introduces several technologies into the standard to improve subjective and objective image quality while showing higher compression ratio than the existing codecs such as MPEG-2 and H.263. One of them is the in-loop deblocking filter, which determines the boundary strength according to the coding conditions generated during video compression encoding and decoding, and applies the normal intensity filter and the strong intensity filter according to the boundary strength. .

In addition, the H.264 / AVC deblocking filter performs adaptive deblocking filtering by dividing the actual edges of the image appearing in the original image frame and the blocking artifacts generated by the encoding and decoding processes in units of blocks. Is implemented.

In this case, by analyzing the pixel data distribution characteristics of adjacent blocks on the basis of the block edge, various types of filter equations are performed in the filter of the normal intensity and the filter of the strong intensity. As such, the H.264 / AVC deblocking filter is highly adaptable and needs to consider complexity and a large amount of computation in actual implementation.

High performance deblocking filter implementation method for high resolution image processing such as HD-TV. In the past, various types of filter equations and computational processes are processed in hardware in parallel, and then appropriately determined according to edge strength coefficients of edges for deblocking filtering. The method of selecting the output was used. However, this approach has the disadvantage of increasing the required hardware area and power consumption.

In addition, there is a method of lowering power by applying a gate clock method to one of a filter of a moderate intensity or a strong intensity filter according to the boundary strength coefficient generated by another conventional method. And it is complicated because all the calculation logic must be implemented in hardware for the various modification modes of the filter.

Accordingly, an object of the present invention is to provide a deblocking filter having a pipelined structure in which a data path for performing intensity strength filtering and a data path for performing strong intensity filtering are integrated. It is done.

A deblocking filtering operation method according to an aspect of the present invention includes: a first step of reading input pixel data to be deblocked filtered and generating an index value necessary for filtering; Generating a threshold value from the generated index value, and reading a boundary strength coefficient for an edge currently being filtered; Determining whether an edge is due to an error due to blocking artifacts using the input pixel data and a threshold value; A fourth step of selectively performing a normal intensity filtering or a strong intensity filtering operation according to the interface intensity coefficient; And a fifth step of outputting the result of the normal intensity filtering or the strong intensity filtering operation.

A step 4-1 of calculating at least one common factor extracted from a plurality of normal intensity filtering or strong intensity filtering equations using the input pixel data; Performing a combination of an addition operation and a shift operation that constitute a normal filtering operation or a strong intensity filtering operation using the calculated at least one common factor; Steps 4-3 to perform a clipping operation according to the normal intensity filtering operation; The method may include performing a fourth operation of adding the clipped value to the pixel data, and the fifth operation may include performing a clipping operation of the result of the fourth operation.

The clipping operation time of the fifth step may be smaller than the clipping operation time of the fourth to third steps.

The at least one common factor may be a common factor that can be calculated by combining two consecutive adders and a shifter.

The result of each step may be stored in pipeline registers provided in stages.

The pipeline register may selectively store the result of the normal strength filtering operation and the strong intensity filtering operation.

The fourth step includes: calculating at least one common factor extracted from a plurality of normal intensity filtering or strong intensity filtering equations using the input pixel data; And performing a plurality of normal intensity filtering or strong intensity filtering equations using the common factor and the input pixel data.

Deblocking filtering operation apparatus according to another aspect of the present invention includes a threshold generator for generating a threshold value from the input index value; An edge determination unit determining whether an edge to be filtered currently is due to an error due to blocking artifacts using the generated threshold value and input pixel data to be filtered; And a filtering operation performing unit for selectively performing a normal intensity filtering or a strong intensity filtering operation according to the input boundary strength coefficient. And an output unit configured to output the result of the normal intensity filtering or the strong intensity filtering operation.

The filtering operation performing unit may include: a plurality of first basic calculation units configured to calculate common factors extracted from a plurality of normal intensity filtering equations and a strong intensity filtering equation using the input pixel data; A plurality of second basic calculation units for performing a normal filtering operation or a strong intensity filtering operation using a plurality of common factors calculated by the plurality of first basic calculation units; A plurality of first clipping operation units for performing a clipping operation according to a normal intensity filtering operation; And a plurality of adders for adding the clipped value with pixel data for an arbitrary cell, and the output unit includes a plurality of second clipping operation units for clipping the added result of the add operator. It can be characterized.

The first basic operation unit and the second basic operation unit may be configured to include two consecutive adders and a shifter.

The threshold generator, the edge determiner, the first basic operation unit, the second basic operation unit, the first clipping operation unit, the plurality of adders and the output unit may be configured to operate as each stage of the pipeline. have.

The apparatus may further include a plurality of pipeline registers for storing a result calculated at each stage of the pipeline. The pipeline register may selectively store the result of the normal intensity filtering operation and the result of the strong intensity filtering operation.

The first basic operation unit and the second basic operation unit may include means for controlling input data to selectively perform one filtering operation among two filtering operations.

A plurality of first basic calculation units for calculating common factors extracted from the plurality of normal intensity filtering equations and the strong intensity filtering equation using the input pixel data; And a plurality of second basic operation units that perform a normal intensity filtering or a strong intensity filtering operation by using the common factor and the input pixel data.

As described above, the deblocking filter including the common data path structure according to the present invention can provide a deblocking filter that has a small area and operates at low power because it uses one integrated data path.

By using the deblocking filter according to the present invention, the blocking artifacts can be eliminated to improve the image quality without loss of compression. In this case, a considerable amount of computation and power are not consumed, which is particularly advantageous for mobile devices.

Hereinafter, an apparatus and method for deblocking filtering including a common data path structure according to the present invention will be described in detail with reference to the accompanying drawings.

In addition, the H.264 / AVC deblocking filter receives up to eight pixel data existing on both sides of an edge between 4 × 4 blocks. Up to six of them are transformed with respect to the coding conditions. The deblocking filter has a high adaptability since the filtering strength is determined according to the coding conditions (eg, interface strength).

The deblocking filter may use interface strength coefficients in connection with coding conditions. First, when the interface strength coefficient is 0, the deblocking filter is not applied. In addition, when the interface strength coefficient is 0 to 3, filtering of the normal intensity is performed, and when the interface strength coefficient is 4, filtering of the strong intensity is performed.

As another coding condition, the deblocking filter compares eight input pixel data consisting of one line with a threshold to determine whether or not it is an actual edge represented in the original image. The deblocking filter may determine a filtering mode by identifying distribution characteristics of pixel data values of the original image.

For the filtering operation according to the coding conditions, the deblocking filter includes all of the filter operation logic of the normal strength, the filter operation logic of the strong strength, and the operation logic of the various deformation modes of the filter. In this case, if the various types of filter equations can be converted into a structure having a common data path and input / output signal mux, the size of the deblocking filter can be significantly reduced, and at the same time, the problem of increased power consumption can be solved.

In particular, in the present invention, the execution conditions of the filter of the normal strength applied when the boundary strength has a value of 0 to 3 and the filter of the strong strength applied when the boundary strength has a value of 4 are mutually exclusive. Exclusive) In this case, the mutually exclusive characteristic means a characteristic in which normal intensity filtering and strong intensity filtering are not simultaneously performed on one pixel line data. Using this feature, the filter of the normal intensity and the strong intensity filter can be merged and processed by the same operation unit.

However, it is not easy to set up common data paths because the filters of normal intensity and strong intensity filter have distinctly different mathematical characteristics. Specifically, a strong intensity filter consists of only an adder and a shift register, while a normal intensity filter mainly consists of a subtractor, and also includes clipping logic that does not exist in a strong intensity filter.

In this case, an Arithmatic Logic Unit (ALU) that can simultaneously perform an adder and a subtractor may be used. However, such an arithmetic logic operator cannot achieve the purpose of miniaturization and low power that the present invention intends to solve because of its complicated structure, large area, and high power consumption.

Hereinafter, a method for performing normal intensity filtering and strong intensity filtering using a common data path will be described. First, for convenience of description, a block for performing edge filtering will be defined.

2 illustrates a block for performing edge filtering according to the present invention.

In FIG. 2A, it can be seen that the left blocks are P blocks and the right block is a Q block with respect to the vertical edge. The pixel data corresponding to the P block to the left based on the edge are represented by p0 to p3 and the pixel data corresponding to the Q block to q0 to q3, respectively.

In FIG. 2B, it can be seen that the upper blocks are P blocks and the lower blocks are Q blocks based on the horizontal edge. Pixel data corresponding to the P block upwards with respect to the edges will be represented by p0 to p3, and pixel data corresponding to the Q block will be represented as q0 to q3, respectively.

In order to perform filtering of the P block and the Q block according to the H.264 / AVC standard, the condition of Equation 1 below must be satisfied. In the equation below, α and β are preset thresholds. The ABS function here means absolute value calculation.

BS! = 0 && (p0-q0) <α && ABS (p1-p0) <β && ABS (q1-q0) <β

When the condition of Equation 1 is satisfied and the boundary strength factor has a value of 0 to 3, normal intensity filtering is applied. The equation in the H.264 / AVC standard for performing normal intensity filtering is shown in Equation 2 below.

Δ = Clip3 (-tc, tc, (((q0-p0) << 2) + (p1-q1) + 4) >> 3))

p`0 = Clip1 (p0 + Δ)

q`0 = Clip1 (q0-Δ)

Here, the clipping coefficient tc is determined differently depending on the case of filtering the edge between the luma 4 × 4 blocks and the case of filtering the edge between the chroma 4 × 4 blocks. The clipping coefficient tc may be obtained as in Equation 3 below.

i) When deblocking filtering edges between luma 4x4 blocks,

tc = tc0 + ((ap <β)? 1: 0) + ((aq <β)? 1: 0)

ii) when deblocking filtering edges between chroma 4x4 blocks,

tc = tc0 + 1

That is, in the filtering of the normal intensity, the filtering operation for P0 and Q0 may be performed by using Equations 2 and 3 below.

On the other hand, the filtering operation may be selectively performed on P1 and Q1 in the normal intensity filtering on the luma 4x4 block. In this case, the equation for filtering the normal strength for P1 and Q1 is shown in Equations 4 and 5 below.

i) when the condition that ABS (p2-p0) <β is satisfied,

p`1 = p1 + Clip3 (-tc0, tc0, (p2 + ((p0 + q0 + 1) >> 1)-(p1 << 1)) >> 1)

ii) if the conditions of i) are not satisfied,

p`1 = p1

i) If ABS (q2-q0) <β is satisfied

q1` = q1 + Clip3 (-tc0, tc0, (q2 + ((p0 + q0 + 1) >> 1)-(q1 << 1)) >> 1)

ii) does not meet the conditions of i)

q'1 = q1

That is, it can be seen that the filtering of the normal intensity for P1 and Q1 is selectively performed depending on whether the ABS (p2-p0) <β condition or the ABS (q2-q0) <β condition is satisfied.

On the other hand, if the boundary strength coefficient (BS) is 4, strong filtering is applied, and the equation in the H.264 / AVC standard for performing such strong filtering is as follows. Unlike normal intensity filtering, strong filtering is performed on up to six pixel data.

Equation 6 below is a strong filtering equation for P0 to P2 blocks based on edges between luma 4x4 blocks.

i) When ABS (p2-p0) <β && ABS (p0-q0) <((α >> 2) + 2)

p'0 = (p2 + 2 * p1 + 2 * p0 + 2 * q0 + q1 + 4) >> 3

p'1 = (p2 + p1 + p0 + q0 + 2) >> 2

p'2 = (2 * p3 + 3 * p2 + p1 + p0 + q0 + 4) >> 3

ii) if the conditions of i) are not satisfied,

p'0 = (2 * p1 + p0 + q1 + 2) >> 2

p'1 = p1

p'2 = p2

Equation 7 below is a strong filtering equation for Q0 to Q2 blocks based on edges between luma 4x4 blocks.

i) When ABS (q2-q0) <β && ABS (p0-q0) <((α >> 2) + 2)

q'0 = (p1 + 2 * p0 + 2 * q0 + 2 * q1 + q2 + 4) >> 3

q'1 = (p0 + q0 + q1 + q2 + 2) >> 2

apply the formula of q'2 = (2 * q3 + 3 * q2 + q1 + q0 + p0 + 4) >> 3,

ii) if the conditions of i) are not satisfied,

q'0 = (2 * q1 + q0 + p1 + 2) >> 2

q'1 = q1

q'2 = q2

In the above, the deblocking filtering method for each condition defined in the H.264 / AVC standard has been described. In the present invention, the above expression is rearranged. Let's take a look at the rearranged formula. The following equations are the result of rearranging the standard filtering algorithm described above, so that given the same filtering conditions and input pixel data, the same expressions are output as in the standard design. In addition, the method of rearranging the formula in the standard may be various modifications, it is noted that the present invention is not limited to the equation described below.

Rearranging Equations 4 and 5 related to the normal intensity filtering applied when the interface intensity coefficient has a value between 0 and 3 is as follows.

i) when ABS (p2-p0) <β is satisfied,

p'1 = p1 + Clip3 (tC0, tC0, (p2 (p1 << 1) + ((p0 + q0 + 1) >> 1)) >> 1)

ii) if the conditions of i) are not satisfied,

p`1 = p1

i) when ABS (q2-q0) <β is satisfied,

q'1 = q1 + Clip3 (tC0, tC0, (q2 (q1 << 1) + ((p0 + q0 + 1) >> 1)) >> 1)

ii) if the conditions of i) are not satisfied,

q`1 = q1

In general, the factors commonly included in Equations 2, 8, and 9, which are necessary for strength filtering, are defined as new variables as shown in Equation 10 below. The variable thus defined is calculated via the basic computing unit of FIG. 3.

((q0-p0) << 2) = N_M0,

(p1-q1) = N_M1,

(p2-(p1 << 1)) = N_L,

(q2-(q1 << 1)) = N_R,

((p0 + q0 + 1) >> 1) = N_SUM

Meanwhile, the formula for calculating a strong filter applied when the boundary strength coefficient BS is 4 may be rearranged as follows to extract common terms. The equation rearranged to extract a common factor for the equation (6) is as shown in the following equation.

i) When ABS (p2-p0) <β && ABS (p0-q0) <((α >> 2) + 2)

p'0 = (2 * (p0 + q0) + (p2 + p1 + 2) + (p1 + q1 + 2)) >> 3

p'1 = ((p0 + q0) + (p2 + p1 + 2)) >> 2

p'2 = ((p0 + q0) + (p2 + p1 + 2) + 2 * (p3 + p2 + 1)) >> 3

ii) if the conditions of i) are not satisfied,

p'0 = ((p1 + p0) + (p1 + q1 + 2)) >> 2

p'1 = p1

p'2 = p2

Equation 11 rearranges Equation 6 in which no common arguments existed, and as a result, common factors such as p0 + q0 or p2 + p1 + 2 exist. Similarly, the equation rearranged to extract common factors for Equation 7 is as follows.

i) When ABS (q2-q0) <β && ABS (p0-q0) <((α >> 2) + 2)

q'0 = (2 * (p0 + q0) + (q2 + q1 + 2) + (p1 + q1 + 2)) >> 3

q'1 = ((p0 + q0) + (q2 + q1 + 2)) >> 2

q'2 = ((p0 + q0) + (q2 + q1 + 2) + 2 * (q3 + q2 + 1)) >> 3

ii) if the conditions of i) are not satisfied,

q'0 = ((q1 + q0) + (p1 + q1 + 2)) >> 2

q'1 = q1

q'2 = q2

Similarly, in Equation 12, it can be seen that there are common factors such as p0 + q0 and q2 + q1 +2.

Here, the common factor included in Equations 11 and 12 for performing strong intensity filtering is defined as a new variable as shown in Equation 13 below. These common factors may be obtained by the basic computing unit of FIG. 3, as in Equation 10.

(p0 + q0) = S_M0,

(p1 + q1 + 2) = S_M1 + 2,

(p2 + p1 + 2) = S_L + 2,

(q1 + q2 + 2) = S_R + 2,

2 * (p3 + p2 + 1) = S_2L + 2,

2 * (q3 + q2 + 1) = S_2R + 2

Here, it can be seen that Equations 10 and 13, which define common factors, can be implemented by up to two addition and shift operations. Equation 13 includes an addition operation having up to three operands and a multiplication operation (shift operation) with two.

In Equation 10, a subtraction operation may be described. However, the subtraction operation can be subtracted using two's complement. In other words, the binary number to be subtracted can be inverted bit by bit, plus 1 to obtain a complement of 2, and added with another operand to perform a subtraction operation. Considering this, Equation 10 can also be expressed through two consecutive addition and shift operations.

This regularity makes it possible for the filters of the normal strength using the subtractor and the strong strength filter composed only of the adder to share the data path as much as possible in the calculation process and to improve the efficiency of the pipeline structure. In addition, since the operation result value is designed to be used as a common intermediate operation value in various equations, the calculation amount is reduced than when each calculation of all the equations is performed.

3 and 4 are diagrams showing the configuration of the basic operation unit for the deblocking filtering operation according to an embodiment of the present invention.

Referring to FIG. 3, the basic computing unit includes a pipeline register, two consecutive adders, a plurality of shifters, a plurality of muxes, and the like. The basic operation unit receives the data stored in the input pipeline register, performs addition and shift operations, and outputs the result to the output pipeline register.

In this case, if the line pixel data p0 to p3 and q0 to q3 are properly input from the input pipeline register to the basic operation unit, the common factors described in Equation 10 or 13 may be calculated.

In addition, when the common argument temporarily stored in the pipeline register is properly input to the basic operation unit, Equations 2, 8, 9, 10, and 11 may be calculated.

In particular, the basic computing unit of FIG. 3 is another feature that utilizes the mutually exclusive features of normal intensity filtering and strong intensity filtering. That is, the basic calculation unit can selectively calculate one of the normal intensity filtering equations (Equations 8 and 9) and one of the strong intensity filtering equations (Equations 0 and 11) at the same time. By controlling a signal input to the basic operation unit through the input / output mux, one of a normal intensity filtering operation or a strong intensity filtering operation may be selectively performed.

For example, the underlined letters in the box on the right side of FIG. 4 mean strong intensity filtering operations, and the dark, italic letters mean normal intensity filtering operations. The basic arithmetic unit of FIG. 4 includes (q2 (q1 << 1) + ((p0) of the condition q q2 of Equation 10 during the strong intensity filtering operation and ② i) of Equation 9 during the filtering operation of the normal intensity. + q0 + 1) >> 1) The part can be selectively operated.

First, in relation to the strong strength filtering operation, the basic operation unit receives S_2R + 2, S_M0, S_R + 2 from the pipeline register. For 2 * (q3 + q2 + 1)), S_2R + 2 is input to the shifter. On the other hand, S_M0 and S_R + 2 are input to the first adder for the term (p0 + q0) + (q2 + q1 + 2). The output of the shifter and the output of the first adder are input to the second adder for the operation of q'2. The operation result of this basic operation unit is the same as the q`2 operation result defined in Equation 10.

In general strength filtering operation of ②, N_SUM_2 is a shifter, and N_L and 0 signals are input to the first adder. The outputs of the shifter and the first adder are likewise input to the second adder. It can be seen that the operation result of the basic operation unit is (q2 (q1 << 1) + ((p0 + q0 + 1) >> 1) which is a part of condition i) of Equation (9).

As a result, one basic operation unit illustrated in FIG. 3 may select two terms corresponding to a filter having a normal intensity and a term corresponding to a filter having a strong intensity one by one to perform a total of two operations. In addition, one basic calculation unit illustrated in FIG. 3 may be implemented to perform two expressions that are not performed simultaneously among the strong filter expressions.

In addition, it is preferable that a plurality of basic arithmetic unit input and output signals are connected to pipeline registers of each stage and constitute a pipeline. In the deblocking filter, a calculation process corresponding to a moderate intensity filter and a strong intensity filter may be performed independently of other pipeline stages according to the edge coefficient of the edge. Thus, in some of the pipeline stages, it is possible to perform a filter operation of moderate intensity, the other part of a strong intensity, and perform the corresponding filter operation in a pipeline structure.

5 is a view showing each step of the pipeline according to another embodiment of the present invention.

In the MR (Memory Read) step of FIG. 5, input pixel data to be deblocked and filtered is read, and α and β, which are index values necessary for filtering, are generated.

In the PRE step, α and β threshold values are generated using the index values generated in the MR step, and the boundary strength coefficients of the edges to be filtered are read.

In the ZERO step, it is determined whether the edge of the image is an actual edge of the original image or an edge to be deblocked. Specifically, in the ZERO step, input pixel line data P0, P1, P2, Q0, Q1, and Q2 are received. This allows | p0-q0 |, | p1-p0 |, | q1-q0 |, | p2-p0 |, | q2-q0 | And compare them with predetermined thresholds.

In the ONE, TWO, THREE, FOUR, and WB (Write Back) phases, basic computational units are placed to control the input and output of the same data path to the MUX, depending on whether to apply normal or strong intensity filters. The deblocking filtering operation is performed.

In the ONE step, the common arguments defined in Equations 10 and 13 are calculated by receiving pixel data of p0 to p3 and q0 to q3. In particular, the basic arithmetic unit illustrated in step 1 of FIG. 5 receives a signal expressed in underlined letters or bold and italic letters, which are signals input during strong intensity filtering and signals for normal intensity filtering, respectively. I looked at it.

Thereafter, in the TWO step, equations included in equations (2), (8), (9), (10), (11), and the like are calculated using the common factors obtained in the ONE step. In this case, the normal intensity filtering is not performed at the TWO stage because clipping operations are included.

That is, in the THREE step, the clipping operation of the normal intensity filter is performed. The clipping operation performed in the THREE step is a process of obtaining Δ of Equation 2. In addition, operations performed in the FOUR step are p0 + Δ and q0-Δ operations belonging to equation (2).

At this time, in the WB (Write Back) step, the clipping operation of the normal intensity filter, that is, the p`0 = Clip1 (p0 + Δ) or q`0 = Clip1 (q0-Δ) operation of Equation 2 is performed. However, since the WB operation includes only a simple clipping operation, it takes much less time than the clipping operation in the THREE step. Therefore, in the WB step, an operation of outputting the result along with the simple clipping operation may be performed.

As can be seen from the above, the normal intensity filtering operation includes a clipping operation, and thus it takes longer than the strong intensity filter operation. Therefore, it is preferable to perform a part of the strong intensity filter operation after the THREE pipeline stage in which the clipping operation is performed (FOUR stage in FIG. 4).

In this case, the intermediate calculation result values required as inputs for the strong intensity filter operation must be passed from the pipeline transfers to the pipeline end through the pipeline registers, which requires additional pipeline registers. Therefore, the area increase due to the registers may be accompanied.

In the present invention, in order to solve this problem, the pipeline register is shared and used. The pipeline registers used to perform normal strength filtering and the pipeline registers used to perform strong strength filtering can be shared. This is because the running conditions of the normal intensity filter and the strong intensity filter are mutually exclusive.

Therefore, at every stage of the pipeline, the result of the combinational logic of each pipeline stage is stored in one shared pipeline register and used as input to the next pipeline stage combination circuit.

In particular, when passing intermediate calculation result values for strong filtering to the rear end of the pipeline, the pipeline register that stores the input values for clipping and the pipeline register that stores the values after clipping It is preferable to use.

In addition, when the boundary strength coefficient is 0 in the deblocking filtering process, input pixel data should be transmitted as output pixel data as it is. When the deblocking filter is designed in a pipeline structure, input pixel data continues to the pipeline every clock. Is entered. In each stage of the pipeline, the input pixel data received in the first stage must be passed in order so that the final stage can be output in a pipelined manner. Therefore, each stage of the pipeline requires a pipeline register for the pixel data.

In the present invention, in order to easily implement the forwarding logic, the result pixel data filtered in the steps before the last step are generated in turn. These result pixel data must also be propagated along the pipeline step. do.

In the present invention, a pipeline register for pixel data propagation is shared with a register for output pixel data propagation by determining a filtering condition and overwriting the resultant pixel data value generated at each stage of the pipeline with a register for input pixel data. How to do it.

Although the present invention has been described in detail through the representative embodiments, those skilled in the art to which the present invention pertains can make various modifications without departing from the scope of the present invention. Will understand. Therefore, the scope of the present invention should not be limited to the described embodiments, but should be defined by the claims below and equivalents thereof.

1 is a block diagram of a video codec encoder in units of typical macroblocks.

2 illustrates a block for performing edge filtering according to the present invention.

3 and 4 are views showing the configuration of a basic operation unit for the deblocking filtering operation according to an embodiment of the present invention.

5 is a diagram illustrating each step operation of a pipeline according to another embodiment of the present invention.

Description of the Related Art [0002]

100: video codec encoder

110: DCT (Discrete Cosine Transform) part

120: quantization unit

130: the motion prediction unit

140: motion compensation unit

150: entropy encoding and decoding unit

160: inverse quantization unit

170: reverse DCT section

180: deblocking filter unit

Claims (15)

In the deblocking filtering operation method, A first step of reading input pixel data to be deblocked filtered and generating an index value necessary for filtering; Generating a threshold value from the generated index value, and reading a boundary strength coefficient for an edge currently being filtered; Determining whether an edge is due to an error due to blocking artifacts using the input pixel data and a threshold value; A fourth step of selectively performing a normal intensity filtering or a strong intensity filtering operation according to the interface intensity coefficient; And And a fifth step of outputting the result of the normal intensity filtering or the strong intensity filtering operation. The method of claim 1, The fourth step, A step 4-1 of calculating at least one common factor extracted from a plurality of normal intensity filtering or strong intensity filtering equations using the input pixel data; Performing a combination of an addition operation and a shift operation constituting a normal filtering operation or a strong intensity filtering operation using the calculated at least one common factor; Steps 4-3 to perform a clipping operation according to the normal intensity filtering operation; And adding a fourth operation of adding the clipped value to pixel data. And the fifth step includes clipping the result of steps 4-4. The method of claim 2, The clipping operation time of the fifth step is smaller than the clipping operation time of the fourth to third steps. The method of claim 2, The at least one common factor is A deblocking filtering method, characterized in that it is a common factor that can be calculated by combining two successive adders and a shifter. The method of claim 2, And the result of each step is stored in a pipeline register provided in stages. The method of claim 5, And the pipeline register selectively stores the result of the normal strength filtering operation and the result of the strong intensity filtering operation. The method of claim 1, The fourth step, Calculating at least one common factor extracted from a plurality of normal intensity filtering or strong intensity filtering equations using the input pixel data; Wow And performing a plurality of normal intensity filtering or strong intensity filtering equations using the common factor and the input pixel data. In the deblocking filtering operation device, A threshold generator for generating a threshold value from the input index value; An edge determination unit determining whether an edge to be filtered currently is due to an error due to blocking artifacts using the generated threshold value and input pixel data to be filtered; A filtering operation performing unit for selectively performing a normal intensity filtering or a strong intensity filtering operation according to an input boundary strength coefficient; And And an output unit for outputting the result of the normal intensity filtering or the strong intensity filtering operation. The method of claim 8, The filtering operation execution unit, A plurality of first basic calculation units for calculating common factors extracted from the plurality of normal intensity filtering equations and the strong intensity filtering equation using the input pixel data; A plurality of second basic calculation units for performing a normal filtering operation or a strong intensity filtering operation using a plurality of common factors calculated by the plurality of first basic calculation units; A plurality of first clipping operation units for performing a clipping operation according to a normal intensity filtering operation; And A plurality of add operators that add the clipped values with pixel data for any cell; And the output unit includes a plurality of second clipping operation units which perform a clipping operation on the result added by the add operator. 10. The method of claim 9, The first basic calculation unit and the second basic calculation unit, A deblocking filtering computing device comprising two successive adders and a shifter. 10. The method of claim 9, The threshold value generator, the edge determination unit, the first basic operation unit, the second basic operation unit, the first clipping operation unit, the plurality of add operation units and the output unit are configured to operate as respective stages of the pipeline. Blocking filtering computing device. The method of claim 11, And a plurality of pipeline registers for storing the result calculated at each stage of the pipeline. The method of claim 12, And the pipeline register selectively stores the result of the normal strength filtering operation and the result of the strong intensity filtering operation. 10. The method of claim 9, The first basic calculation unit and the second basic calculation unit, And means for controlling input data to selectively perform one filtering operation of two filtering operations. The method of claim 8, A plurality of first basic calculation units for calculating common factors extracted from the plurality of normal intensity filtering equations and the strong intensity filtering equation using the input pixel data; and And a plurality of second basic computing units for performing a normal intensity filtering or a strong intensity filtering operation using the common factor and the input pixel data.

Images