Information processor and inter-prediction mode determination method

The present invention relates to a moving image coding technique suitable for application to an information processing device such as a personal computer.

In recent years, personal computers equipped with software encoders that encode moving images with software have begun to spread. Recently, the H.264 / AVC (Advanced Video Coding) standard has been attracting attention as a next-generation video compression coding technology. This H.264 / AVC standard is a compression coding technology that is more efficient than conventional compression coding technologies such as MPEG2 and MPEG4. Therefore, the encoding process corresponding to the H.264 / AVC standard requires a larger amount of processing than the conventional compression coding technology such as MPEG2 and MPEG4. For this reason, various proposals and the like for reducing the amount of moving image encoding processing have been made so far (see, for example, Patent Document 1 and the like).<patcit num="1"><text>Japanese Unexamined Patent Publication No. 2006-25077</text></patcit>

<p> In the encoding process corresponding to the H.264 / AVC standard, the amount of processing for determining the prediction mode for each macroblock is large. In particular, in high profile (HP: High Profile), in inter-prediction, when the block size of the prediction mode is 8 pixels x 8 pixels or more, the block size of 4 pixels x 4 pixels DCT (Discrete Cosine Transform), 8 pixels x Since it is possible to select the optimum one from the DCTs with a block size of 8 pixels (substantially, there are prediction modes of the number of prediction modes x 2), the prediction mode determination of the inter-prediction is made. The amount of processing required for this increases in proportion to the number of prediction mode candidates. Therefore, for example, a mechanism for efficiently performing this prediction mode determination while suppressing deterioration of image quality is strongly desired.</p><p> The present invention has been made in consideration of such circumstances, and an object of the present invention is to provide an information processing apparatus and an inter-prediction mode determination method capable of efficiently performing an inter-prediction prediction mode determination. To do.</p>

<p> In order to achieve the above-mentioned object, the information processing apparatus of the present invention is an information processing apparatus that encodes a moving image signal, and processes macroblocks obtained by dividing a screen to be encoded with different block sizes. Between an inter-prediction means that generates a prediction signal using any of the inter-prediction modes among a plurality of types of inter-prediction modes as a unit, and a prediction signal generated by the inter-prediction means and an image signal of the macroblock. Multiple types of DCT (Discrete Cosine) with different block sizes as the processing unit for the predicted residual signal of The combination of the conversion means for orthogonal transform using any of the DCTs in Transform), the inter-prediction mode used by the inter-prediction means, and the DCT used by the conversion means for each macroblock is of the plurality of types. The inter-prediction mode determining means for determining from the inter-predicting mode and the plurality of types of DCT is provided, and the inter-predicting mode determining means targets a specific type of DCT among the plurality of types of DCT. A first selection means for selecting a predetermined number of inter-prediction modes from the plurality of types of inter-prediction modes, the predetermined number of inter-prediction modes selected by the first selection means, and the plurality of types of DCTs. It is characterized by having one inter-prediction mode and a second selection means for selecting a combination of DCTs.</p><p> Further, in the inter-prediction mode determination method of the present invention, one of a plurality of types of inter-prediction modes having different block sizes as processing units is used for a macro block obtained by dividing a screen to be encoded. A plurality of types of inter-prediction means for generating a prediction signal using the inter-prediction means, and a plurality of types of prediction residual signals between the prediction signal generated by the inter-prediction means and the image signal of the macroblock, with different block sizes as processing units. DCT (Discrete Cosine) It is an inter-prediction mode determination method of an information processing apparatus that includes a conversion means for orthogonal transform using any of the DCTs in Transform) and encodes a moving image signal. Targeting a specific type of DCT among the types of DCT, a predetermined number of inter-prediction modes are selected from the plurality of types of inter-prediction modes, and the selected predetermined number of inter-prediction modes and the plurality of types of DCT are selected. Select one combination of the inter-prediction mode and the DCT from the above, and use this combination as the inter-prediction mode and the conversion used by the inter-prediction means in the plurality of types of inter-prediction modes and the plurality of types of DCTs. It is characterized in that it is determined that it is a combination with the DCT used by the means.</p>

<p> According to the present invention, it is possible to provide an information processing device and an inter-prediction mode determination method that enable efficient inter-prediction prediction mode determination.</p>

Hereinafter, an embodiment of the present invention will be described with reference to the drawings. FIG. 1 shows a configuration example of the information processing apparatus according to the present embodiment. This information processing device is realized as, for example, a battery-powered notebook-type personal computer or the like.

As shown in Figure 1, this computer has CPU 11, Northbridge 12, Main Memory 13, Graphics Controller 14, VRAM14A, LCD15, Southbridge 16, BIOS-ROM17, HDD18, HD DVD19, Sound Controller 20, Speaker 21, It is equipped with an embedded controller / keyboard controller IC (EC / KBC) 22, a keyboard 23, a touch pad 24, a power supply circuit 25, a battery 26, a network controller 27, and the like.

The CPU 11 is a processor that controls the operation of each part in the computer. The CPU 11 executes an operating system (OS) 100 loaded from the HDD 18 into the main memory 13 and various application programs including utilities that operate under the control of the OS 100. The video encoder application 200 is included in these various application programs. The video encoder application 200 is software for encoding a moving image, and operates as a software encoder corresponding to the H.264 / AVC standard. CPU 11 also executes the BIOS stored in BIOS-ROM17. The BIOS is a program for controlling various hardware.

The north bridge 12 is a bridge device that connects the local bus of CPU 11 and the south bridge 16. The north bridge 12 has a function of executing communication with the graphics controller 14 via a bus, and also has a built-in memory controller that controls access to the main memory 13. The graphics controller 14 is a display controller that controls an LCD 15 used as a display monitor of the present computer. The graphics controller 14 generates a display signal to be sent to the LCD 15 from the image data written in the VRAM 14A.

The south bridge 16 is a controller that controls various devices on the PCI bus and the LPC bus. In addition, the BIOS-ROM17, HDD18, HD DVD19 and sound controller 20 are directly connected to the south bridge 16 and have a function to control them. The sound controller 20 is a sound source controller that controls the speaker 21.

The EC / KBC22 is a one-chip microcomputer in which an embedded controller for power management and a keyboard controller for controlling a keyboard 23 and a touchpad 24 are integrated. The EC / KBC22 cooperates with the power supply circuit 25 to supply and control power from the battery 26 or an external AC power supply to each part. The network controller 27 is a communication device that executes communication with an external network such as the Internet.

Next, with reference to FIG. 2, the functional configuration of the software encoder realized by the video encoder application 200 operating on the computer having such a hardware configuration will be described.

The encoding process by the video encoder application 200 corresponds to the H.264 / AVC standard, and as shown in the figure, the video encoder application 200 has an input unit 201, a DCT / quantization unit 202, an entropy coding unit 203, and a reverse. It is equipped with a quantization / inverse DCT unit 204, an intra prediction unit 205, a deblocking filter 206, a frame memory 207, a motion detection unit 208, an inter prediction unit 209, a prediction mode determination unit 210, a calculator 211, 212, and the like.

The video encoder application 200 encodes each screen (picture) input from the input unit 201 in units of macroblocks of, for example, 16 × 16 pixels. The prediction mode determination unit 210 selects either the intra-frame prediction coding mode (intra prediction mode) or the motion compensation inter-frame prediction coding mode (inter prediction mode) for each macroblock. FIG. 3 shows a functional block of the prediction mode determination unit 210.

As shown in FIG. 3, the prediction mode determination unit 210 includes an intra prediction mode determination unit 2101, an inter prediction mode determination unit 2102, and an intra inter prediction mode determination unit 2103.

In both the intra prediction mode and the inter prediction mode, there are a plurality of prediction mode candidates that can be selected for each macroblock. First, each of the intra prediction mode determination unit 2101 and the inter prediction mode determination unit 2102. However, the most cost-effective (small code amount) prediction mode candidate is selected from the plurality of prediction mode candidates. Secondly, the intra-inter-prediction mode determination unit 2103 compares the two prediction mode candidates selected by the intra-prediction mode determination unit 2101 and the inter-prediction mode determination unit 2102, respectively, and has a higher cost. The prediction mode, that is, either the intra prediction mode or the inter prediction mode, is finally selected.

By the way, in the high profile of the H.264 / AVC standard, in the inter-prediction, when the block size of the prediction mode is 8 pixels x 8 pixels or more, the DCT of the block size of 4 pixels x 4 pixels, the block of 8 pixels x 8 pixels It is possible to select the most suitable one from the DCTs of the size. Therefore, if there are m types of prediction modes and n types of DCT can be selected in a certain standard, as shown in FIG. 4, m × n prediction mode candidates can be selected only by inter-prediction. It will actually exist. In such a situation, if all prediction mode candidates are evaluated and the optimum prediction mode candidate is selected without any ingenuity, the amount of processing required for the prediction mode determination of the inter-prediction becomes enormous. It ends up. Therefore, the inter-prediction mode determination unit 2102 of the present computer is provided with a mechanism for efficiently performing this prediction mode determination while suppressing deterioration of image quality, for example, and this point will be described in detail below.

In the software encoder whose functional configuration is shown in FIG. 2, the intra prediction unit 205 generates the prediction signal s1 from the coding target screen (picture) in the intra prediction coding mode, and the coding target screen (picture). The prediction error signal s2 obtained by subtracting this prediction signal s1 from) is orthogonally converted and quantized by the DCT / quantization unit 202, and the entropy coding unit 203 is subjected to the intra prediction mode information and the quantized orthogonal transform coefficient. It is encoded by performing entropy coding.

On the other hand, in the inter-prediction coding mode, the motion detection unit 208 first estimates the motion from the already encoded screen (picture) stored in the frame memory 207, and then the inter-prediction unit 209 , The motion compensation frame-to-frame prediction signal s3 corresponding to the coded target screen is generated in the specified shape unit. Then, the DCT / quantization unit 202 orthogonally transforms and quantizes the prediction error signal s4 obtained by subtracting the motion compensation frame-to-frame prediction signal s3 from the coded target screen (picture), and the entropy coding unit 203 performs the inter-prediction mode. The information and the quantized discrete cosine transform coefficient are encoded by entropy coding.

In addition, the inverse quantization / inverse DCT section 204 reverse-quantizes and inverse-orthogonally transforms the quantization coefficient of the orthogonal transform and the quantized image (picture), and the deblocking filter 206 reduces block noise. Deblocking filter processing is performed to perform this.

FIG. 5 is a diagram for explaining the basic principle of the prediction mode determination executed by the inter-prediction mode determination unit 2102 of the present computer.

Candidates for the optimum prediction mode have the property of being the same at a certain rate regardless of the DCT. Therefore, even when the prediction mode is selected first and the prediction mode for determining the optimum DCT selection is narrowed down, it can be said that the ratio of selecting the truly optimum prediction mode and the combination of the DCT is considerably high. Focusing on this property, the inter-prediction mode determination unit 2102 of this computer first determines the prediction mode under specific DCT conditions and selects the optimum prediction mode candidate (Fig. 5 (1)). Then, the inter-prediction mode determination unit 2102 determines to select the optimum DCT only for the optimum prediction mode candidate, and finally determines the prediction mode (FIG. 5 (2)). At this time, since the specific DCT has already been evaluated when the optimum prediction mode candidate is selected, it is sufficient to evaluate only the DCT other than the specific DCT.

To explain with a more specific example, in the above-mentioned H.264 / AVC standard high profile, as shown in FIG. 6, 16 pixels × 16 pixels, 16 pixels × 8 pixels, 8 pixels × 16 pixels, 8 There are four types of prediction modes in which different block sizes of pixels x 8 pixels are used as processing units. Further, in this H.264 / AVC standard high profile, as shown in FIG. 7, there are two types of DCTs having different block sizes of 4 pixels × 4 pixels and 8 pixels × 8 pixels as processing units. Therefore, the number of prediction mode candidates is 4 (m in FIGS. 4 and 5) × 2 (n in FIGS. 4 and 5) = 8.

On the other hand, the inter-prediction mode determination unit 2102 of this computer first has one type (N in FIG. 5) of, for example, 4 pixels × 4 pixels among two types of DCTs of 4 pixels × 4 pixels and 8 pixels × 8 pixels. Evaluate the prediction mode using only DCT (the number of candidates is 4 (m) x 1 (N)), and select, for example, 1 (M in Fig. 5) optimal prediction mode candidates. Then, only one optimum prediction mode candidate obtained by this 4-pixel x 4-pixel DCT is evaluated by the DCT of the other 8-pixel x 8 image (the number of candidates is 1 (M)). × (2 (n) -1 (N))), finally determine the prediction mode.

In other words, in this case (with one optimal prediction mode candidate), the inter-prediction mode determination unit 2102 of this computer realizes that the number of prediction mode candidates can be reduced to 4 + 1 = 5 (three). .. Since there is a high probability that the optimum prediction mode will be the same for the DCT of 4 pixels x 4 pixels and the DCT of 8 pixels x 8 pixels, the ratio of selecting the truly optimum combination is high, and there is almost no deterioration in image quality.

Further, the inter-prediction mode determination unit 2102 of the present computer controls that the optimum prediction mode candidate is set to one in the B picture, which is easy to predict, and is increased to 2 in the P picture. That is, the inter-prediction mode determination unit 2102 determines the number of prediction mode candidates according to the type of picture, thereby reducing the amount of processing required for prediction mode determination while adaptively suppressing image quality deterioration. There is.

FIG. 8 is a flowchart showing an operation procedure of the prediction mode determination executed by the inter-prediction mode determination unit 2102 of the present computer.

The inter-prediction mode determination unit 2102 first determines whether the screen to be encoded is a P picture or a B picture (step A1), and if it is a P picture, it is 2, and if it is a B picture, it is 1. Based on the judgment result, the number of optimum prediction mode candidates is determined (step A2).

Next, the inter-prediction mode determination unit 2102 calculates the cost of each of all types of prediction modes for a specific type of DCT among the plurality of types of DCTs (step A3). Based on the result of this cost calculation, the inter-prediction mode determination unit 2102 selects the optimum prediction mode candidates for the number of previously determined candidates (step A4).

Then, the inter-prediction mode determination unit 2102 calculates the cost of each of the optimum prediction mode candidates selected by the number of previously determined candidates for the DCTs other than the specific type DCT (step A5), (step A3). Determine the optimal prediction mode, DCT (from among the costs associated with a particular type of DCT calculated in) (step A6).

As described above, according to this computer, it is possible to efficiently perform the prediction mode determination of inter-prediction without causing deterioration of image quality, for example.

The present invention is not limited to the above-described embodiment as it is, and at the implementation stage, the components can be modified and embodied within a range that does not deviate from the gist thereof. In addition, various inventions can be formed by an appropriate combination of the plurality of components disclosed in the above-described embodiment. For example, some components may be removed from all the components shown in the embodiments. In addition, components across different embodiments may be combined as appropriate.

<figref num="1">The figure which shows the structural example of the information processing apparatus (personal computer) which concerns on one Embodiment of this invention.</figref><figref num="2">The figure which shows the functional structure of the software encoder realized by the video encoder application running on the computer of the same embodiment.</figref><figref num="3">The figure which shows the functional block of the prediction mode determination part of the video encoder application which operates on the computer of the same embodiment.</figref><figref num="4">Diagram for explaining the general basic principle of prediction mode judgment of inter-prediction</figref><figref num="5">The figure for demonstrating the basic principle of the prediction mode judgment executed by the inter prediction mode judgment part of the video encoder application operating on the computer of the same embodiment.</figref><figref num="6">FIG. 1 for explaining a specific example of prediction mode determination executed by the inter-prediction mode determination unit of the video encoder application operating on the computer of the same embodiment (a diagram illustrating the types of prediction modes).</figref><figref num="7">A second figure for explaining a specific example of the prediction mode judgment executed by the inter-prediction mode judgment unit of the video encoder application operating on the computer of the same embodiment (a figure illustrating the type of DCT).</figref><figref num="8">A flowchart showing the operation procedure of the prediction mode determination executed by the inter-prediction mode determination unit of the video encoder application operating on the computer of the same embodiment.</figref>

Code description

11 ... CPU, 12 ... Northbridge, 13 ... Main memory, 14 ... Graphics controller, 14A ... VRAM, 15 ... LCD, 16 ... Southbridge, 17.. .BIOS-ROM, 18 ... HDD, 19 ... HD DVD, 20 ... Sound Controller, 21 ... Speaker, 22 ... Embedded Controller / Keyboard Controller IC (EC / KBC), 23 ... Keyboard, 24 ... Touch Pad, 25 ... Power Circuit , 26 ... Battery, 27 ... Network Controller, 100 ... Operating System (OS), 200 ... Video Encoder Application, 201 ... Input, 202 ... DCT / Quantization, 203 ... Entropy coding section, 204 ... Inverse quantization / inverse DCT section, 205 ... Intra prediction section, 206 ... Deblocking filter, 207 ... Frame memory, 208 ... Motion detection section , 209 ... Inter Prediction Mode, 210 ... Prediction Mode Judgment, 211, 212 ... Calculater, 2101 ... Intra Prediction Mode Judgment, 2102 ... Inter Prediction Mode Judgment, 2103 ... Intra -Inter prediction mode judgment unit.