KR20040039809A - Moving picture encoder and method for coding using the same - Google Patents

Abstract

PURPOSE: A moving picture encoder and a coding method using the same are provided to reduce VLC(Variable Length Coding) calculation quantity by detecting scene change or complex motion using an SAD(Sum Absolute Difference) value obtained for motion compensation and selecting whether using an intra VLC table or an inter VLC table. CONSTITUTION: A DCT(Discrete Cosine Transform) unit(11) performs the DCT of an image signal provided from the outside. A quantization unit(13) quantizes and samples the image signal transformed in the DCT unit(11). An inverse quantization unit(15) and an inverse DCT unit(17) recover the image coding signal quantized in the quantization unit(13). An image memory(18) stores the image recovered in the inverse DCT unit(17) for using the stored image as a reference image for the motion compensation of an inputted image. A motion predictor(19) estimates motion with reference to the image stored in the image memory(18). A VLC mode selector(14) compares an SAD value of the macro block predicted in the motion predictor(19) with an SAD value of a previous macro block, and selects whether using an inter VLC table or an intra VLC table. A VLC unit(20) performs VLC according to the VLC table selected in the VLC mode selector(14).

Description

MOVING PICTURE ENCODER AND METHOD FOR CODING USING THE SAME}

The present invention relates to a video encoder and an encoding method using the same. More particularly, when encoding for motion estimation, a video encoder having improved coding efficiency by reducing the amount of VLC calculation according to image characteristics and encoding using the same are provided. It is about a method.

The current and future communication environment is rapidly changing beyond the distinction between wired and wireless areas and regional countries. In particular, future communication environments, such as IMT-2000, require users as well as video and audio. The trend is to build an environment that provides various information in real time or comprehensively.

In addition, the development of the personal mobile communication system is not only a voice communication in the cellular phone or PCS, but also to transmit text information as well as to access the Internet wirelessly using a personal mobile communication terminal, or to watch videos on TV It is being developed to transmit the data.

In particular, a digital television system that processes a video into digital data, transmits the data in real time, receives the video, and displays the video, and the video transmitted in real time is becoming an essential element in a personal portable terminal using the IMT2000.

In the related art, the portable terminal transmits and receives only human voice. However, due to the development of multimedia and the development of digital information processing technology, various information such as voice and video can be transmitted.

Above all, this technology has been commercialized. First of all, the analog video signal is subjected to special digital processing such as quantization and variable length encoding, and then included in the digital information and transmitted. It greatly contributed to the development of video compression technology that allows abundant information transmission and reception.

Recently, with the development of digital signal processing technology, methods of compressing and transmitting a large amount of video information through a limited bandwidth transmission channel have been developed. However, if an error occurs on the transmission channel, the quality of the reconstructed image is greatly reduced. Occurs.

At this time, error concealment techniques have been studied to recover near to the original image by supplementing the lost information from the neighboring information that is normally restored without using an error correction code to maximize the limited bandwidth.

In particular, in the case of an MPEG (Moving Picture Experts Group), if an error occurs, all information until the next sync signal is found is lost. In addition, since the motion compensation encoding technique is used, the influence of the damaged portion is continued over the subsequent frames.

The present invention is an IMT-2000 video terminal that applies an optimized motion decision value and an intra mode compression algorithm to a hardware accelerator structure that enables fast encoding and quality improvement of video data. Motion decision value of MPEG compression method, which is generally used for wireless image communication, is the most important because it has the largest amount of computation and directly affects the quality and the amount of bits.

Therefore, many studies on motion judgments have been made continuously, and recently, the motion judgments have the same image quality and bit rate as the full search using the diamond search pattern, and are most commonly used. A three step search method is used.

1 is a configuration diagram of a codec for video encoding and decoding which is generally used.

As shown in FIG. 1, when a video signal photographing a real object enters the IMT-2000 terminal 10, an encoding operation is performed to digitally convert it in the encoder 3 and then compress it into an image bitstream. . Therefore, the source coder 7 codes the video, the sound, and the like included in the video signal together with the coding controller 5 according to a coding method suitable for each method.

The source coder 7 and the coding controller 5 in the encoder 3 are separated from each other by coding the video signal. When various signals such as video and sound are coded by the coding controller 5, the signals are separated from the video multiflux coder 8, and then the externally coded video bitstream is transmitted through the transmission buffer 9.

The coding controller 5 of the encoder 3 includes a discrete cosine transform (DCT) for encoding a video signal, a quantization unit for quantizing a video signal converted into a frequency form, and a VLC (Variable Length) for entropy a quantized video signal. Coding), a motion predictor, an inverse quantization unit, an inverse DCT, and an image memory to predict an image signal and reduce coding encoding.

In this case, in order to insert an option mode added to the image coded by the encoder 3, the option selected by the option selector is applied to the IMT-2000 terminal system 10, and the encoder 3 is applied according to the applied option information. Code the video signal.

In addition, the IMT-2000 terminal system 10 includes a decoder 15 for restoring an image bit stream transmitted from the outside, and the decoder 15 temporarily stores the image bit stream transmitted from the outside. The reception buffer 13 and the video multiplex decoder 12 for separating the video signal and the audio signal from the video bitstream stored in the reception buffer 13 are arranged.

The video and audio signals separated by the video multiplex decoder 12 are reconstructed by the source decoder 11, and in particular, to recover the video signal, a variable length decoder and an error are used. Error detection and restoration for detection, motion vector reconstruction and image reproduction, inverse quantization for inverse quantization and inverse discrete cosine transformation to reverse the compression scheme performed by the encoder, and inverse DCT Discrete Cosine Transform), Motion Vector Compensation (MC), and a display unit for displaying a video.

In general, a decoder of an IMT-2000 terminal system reconstructs an image by decoding the image through a process opposite to that encoded by the encoder.

In the codec having the structure described above, when the ANNEX S option is selected from the option selector, variable length encoding is used to use the intra VLC table when scene transition or complex motion is encoded in the inter mode, thereby improving coding efficiency.

As described above, the method of applying the ANNEX S option is a method of encoding by sending an regulation to use ANNEX S unconditionally for picture header information so that intra VLC is also used in inter mode.

Secondly, when intra VLC uses fewer bits so that both inter VLC and intra VLC tables can be applied, intra VLC is used.

Third, the decoder performs VLC on the video bitstream sent to the counterpart using the inter VLC table, and at this time, the intra VLC is used when the decoder detects that the block goes out of the 64 pixel coefficient ranges.

However, when a scene change or a complex scene appears by the ANNEX S option in the inter mode as described above, the method of using the intra VLC table has the following problem.

First, regardless of the characteristics of the image, the use of the intra VLC table by unconditionally adding the ANNEX S option to the picture header information has a problem in that the coding efficiency is lowered because the intra VLC table is used even if there is no scene change or a simple image.

Second, applying both the inter VLC table and the intra VLC table and then using the VLC table having the lowest bit rate has a problem that the calculation is complicated because VLC must be performed for both inter and intra.

Third, when the other side detects the use of the intra VLC, the encoding method using the VLC table is a method corresponding to an error that may occur during the transmission of a video encoding signal, and thus has a distance from the performance of encoding by an actual encoder.

According to the present invention, when the motion estimation is performed in the inter mode, it is determined whether a large number of scene transitions have occurred in the video block or whether the video block has a high complexity. An object of the present invention is to provide a video encoder and a coding method using the same.

1 is a block diagram of a codec for video encoding and decoding generally used.

2 is a diagram illustrating a structure of a video encoder for encoding a video signal according to the present invention.

3 is a flowchart for explaining an encoding process according to the present invention.

* Description of the symbols for the main parts of the drawings *

11: DCT13: quantization unit

14: VLC mode selector 15: inverse quantizer

17: Inverse DCT18: Picture Memory

19: motion predictor 20: VLC

In order to achieve the above object, the video encoder according to the present invention,

Discrete Cosine Transform (DCT) for performing discrete cosine transform of an image signal applied from the outside;

A quantizer configured to quantize and sample a discrete cosine-converted video signal by the DCT;

An inverse quantizer and an inverse DCT for reconstructing the image coded signal quantized by the quantizer;

An image memory configured to store the image reconstructed in the inverse DCT to be used as a reference image for motion compensation of a later input image;

A motion estimator for performing motion estimation with reference to an image stored in the image memory;

A VLC mode selector for selecting whether to use an inter VLC table or an intra VLC table by comparing the SAD value of the macro block predicted by the motion predictor with the SAD value of the previous macro block; And

And VLC performing variable length coding according to the VLC table selected by the VLC mode selector.

Here, the VLC mode selector stores the previous macro block SAD value in order to compare the SAD value of the macro block for the current motion estimation.

In addition, the video encoding method according to another embodiment of the present invention,

Storing the SAD value of the previous macro block used for motion compensation;

Calculating a SAD value after compensating for the current macro block;

Determining whether the SAD value for the current macroblock is equal to or greater than an α-fold value of the SAD value of the previous macroblock; And

If the SAD value for the current macroblock is equal to or greater than an α-fold value of the SAD value of the previous macroblock, encoding using an intra VLC table; It is characterized by.

Here, the value of α is in the range of 1 to 10, and if the SAD value for the current macro block is smaller than the α times the SAD value of the previous macro block, encoding is continued using the inter VLC table. It is done.

The SAD value of the stored previous macroblock may be used to determine whether to use an inter VLC table and an intra VLC table.

According to the present invention, when performing the motion estimation in the inter mode, the inter- and intra VLC may be selectively determined by determining whether the scene block has a lot of scene change or a complex image block using the SAD value obtained from the DCT. By doing so, there is an advantage that the coding efficiency can be increased while reducing the VLC calculation amount.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

2 is a diagram illustrating a structure of a video encoder for encoding a video signal according to the present invention.

As shown in FIG. 2, the video encoder performs encoding by performing a discrete cosine transform and quantization on a video signal applied from the outside, and reconstructs the quantized video coded signal to reencode the video signal previously applied. In the case of the same signal, encoded data of the transmitted video signal is reduced by using the encoded video.

The video encoder includes a DCT (Discrete Cosine Transform) 11 for performing discrete cosine transform on a video signal transmitted from the outside, and a quantization unit 13 for quantizing the discrete cosine transformed video signal in DCT in frequency form, An inverse quantization unit 15 and an inverse DCT 17 for inverse quantization and inverse DCT transformation to use the quantized image coded signal as a reference image during motion estimation (compensation), and the inverse quantization unit The image memory 18 storing the reference image reconstructed by the inverse DCT and the reference image stored in the image memory when the image signal applied from the outside is in the inter mode are compared to the image block in which the motion is generated. When predicting a motion using the motion predictor 19 for compensation and the motion predictor 19, and quantizing it in the quantization unit 13, The VLC mode selector 14 determines whether to use an inter VLC table and an intra VLC table after determining image characteristics using a Sab (Sum Absolute Difference) value, and a VLC table selected by the VLC mode selector 14. It consists of VLC (Variable Length Coding) 20 which performs the variable length coding process.

Although shown in the figure, 12a and 12b, which are not described, represent adders for summing signals.

The video encoder having the above structure performs discrete cosine transform and quantization when an input video signal is an intra-frame, and then there is no previous reference image for image prediction. Intra mode encoding is performed to predict an image block.

However, when the input video signal is a P picture having motion based on the intra frame-coded video frame, whether to perform intra coding on each block in which motion has occurred or to obtain a motion prediction value is first encoded. The encoding performance is improved by appropriately selecting whether to reproduce an image using blocks in the received image frame.

In particular, the present invention estimates the reference image from the stored image to compensate for the motion of the macroblock in the inter mode, and then compares the SAD values when performing variable length encoding to determine the change and the complexity of the image, thereby selecting it in the inter mode. The intra VLC table can be used to reduce the computation of VLC.

When the macroblock to be encoded is in the inter mode, the motion difference with the current macroblock is obtained based on the reference picture stored in the image memory 18 and then VLC is performed using the inter VLC table.

However, even in the inter mode, image switching and complexity may occur. In this case, when the inter VLC table is used in a batch, the VLC coding efficiency is not only decreased, but the complexity of the calculation amount is increased.

Therefore, the SAD value is stored by using the SAD value for each block constituting the macro block in the motion estimation in the inter mode, and the image conversion and complexity are increased by comparing the SAD value of the macro block for the current motion estimation. Judge.

In the inter mode, since the encoding efficiency is increased by determining the degree of motion with the currently encoded image based on the reference image and using the previously encoded image macroblock as it is, the SAD value of the reference image and the macroblock currently encoded. Obtain

Since the encoding is performed in units of the macroblock, the SAD value is stored for each block constituting the macroblock, and then the error value is encoded after the motion estimation for the current macroblock. You will know.

The quantization unit obtains the stored SAD value of the previous macro block and the current motion compensation for the image blocks quantized by the inter mode so that the inter VLC table or the intra VLC table can be selectively used in the VLC process. When the SAD value is suddenly increased by comparing the SAD value, it is determined that the scene of the image is changed or the complexity is increased.

In the VLC mode selector, if the SAD values for the current macroblock obtained after the motion estimation are α times or more than the previously stored SAD values, the intra VLC table is used for the variable length coding even if the motion estimation is performed in the inter mode. VLC was used.

The range of α value has a value in the range of 1 to 10 as a constant and can be determined selectively.

Therefore, the inter VLC table is reduced and the coding efficiency is increased by comparing the SAD values in the inter mode and using the intra VLC table when the scene conversion or the image complexity increases, such as the intra mode.

3 is a flowchart for explaining an encoding process according to the present invention.

As shown in FIG. 3, the video encoder selects an inter mode and an intra mode. In the inter mode, the video encoding is performed by performing motion estimation on the basis of determining that temporal movement has occurred from a previously encoded reference picture, thereby increasing encoding efficiency. I was.

That is, the motion position of the previously encoded image is determined to use the previously encoded image.

At this time, after compensating for the previous macroblocks in the motion compensation step for variable length encoding, the obtained SAD value is stored (S301), and the error is estimated after motion estimation for the macroblock in which the current encoding is performed. The value is compared with the current SAD value obtained when the value is encoded (S302, S303).

When comparing the SAD value for the current macro block with the previously stored SAD value, if the SAD value for the current macro block is greater than or equal to an α-fold value of the SAD value of the previous macro block, the SAD value is severely changed and the image is displayed. It is determined that the conversion or complexity has increased (S303).

Thus, in the inter mode for motion estimation, variable length coding is performed using an intra VLC table to increase the efficiency of VLC.

The range of α values is determined to have a range of 1 to 10, and the user can select a VLC table by arbitrarily selecting a value according to the characteristics of the image (S304).

In addition, when the SAD value for the current macroblock is smaller than α times the SAD value of the previous macroblock, it is determined that the change of the SAD value is not large and variable length encoding is performed using the inter mode VLC table (S305). .

Accordingly, in the present invention, the VLC calculation amount is reduced by selectively applying intra VLC or inter VLC to determine the characteristics of the image by comparing SAD values for each of the blocks necessarily used for motion estimation in the inter mode.

As described in detail above, the present invention has an effect of reducing the VLC calculation amount by selecting whether to use an intra VLC table or an inter VLC table by detecting a scene change or a complex motion by using a SAD value obtained for motion compensation. have.

The present invention is not limited to the above-described embodiments, and various changes can be made by those skilled in the art without departing from the gist of the present invention as claimed in the following claims.

Claims (6)

Discrete Cosine Transform (DCT) for performing discrete cosine transform of an image signal applied from the outside; A quantizer configured to quantize and sample a discrete cosine-converted video signal by the DCT; An inverse quantizer and an inverse DCT for reconstructing the image coded signal quantized by the quantizer; An image memory configured to store the image reconstructed in the inverse DCT to be used as a reference image for motion compensation of a later input image; A motion estimator for performing motion estimation with reference to an image stored in the image memory; A VLC mode selector for selecting whether to use an inter VLC table or an intra VLC table by comparing the SAD value of the macro block predicted by the motion predictor with the SAD value of the previous macro block; And And a VLC for variable length encoding according to the VLC table selected by the VLC mode selector. The method of claim 1, And the VLC mode selector stores a previous macro block SAD value to be compared with an SAD value of a macro block for current motion estimation. Storing the SAD value of the previous macro block used for motion compensation; Calculating a SAD value after compensating for the current macro block; Determining whether the SAD value for the current macroblock is equal to or greater than an α-fold value of the SAD value of the previous macroblock; And If the SAD value for the current macroblock is equal to or greater than an α-fold value of the SAD value of the previous macroblock, encoding using an intra VLC table; Video encoding method characterized in that. The method of claim 3, wherein The range of the α value is a video encoding method, characterized in that 1 to 10 range. The method of claim 3, wherein And if the SAD value for the current macroblock is smaller than the α-fold value of the SAD value of the previous macroblock, encoding is continued using the inter VLC table. The method of claim 3, wherein And a SAD value of the stored previous macroblock is used to determine whether to use an inter VLC table and an intra VLC table.