KR100203719B1 - Method for detecting error of bit stream in mpeg-2 algorithm - Google Patents

Abstract

The present invention relates to an error detection method of a bit stream in which an error in a bit stream of a reconstructed MPEG-2 image frame can be effectively detected by using quantization scale information included in macro block data on a bit stream of an image frame. To this end, the present invention, by inputting the bit stream of the video frame restored to the original signal before being encoded by the decoding means to perform a codeword search for a plurality of pieces of information contained in the macro block continuously Retrieve codewords for quantization scale information of bits; Detecting an information value of a codeword for the retrieved quantization scale information; If it is determined that the detected quantization scale information value is an integer other than zero, no error signal is generated. If it is determined that the detected quantization scale information value is zero, an error signal is generated and provided to the error concealment and recovery means. ; MPEG-2 video bits are detected by performing error detection on the video bit stream by performing error concealment and recovery in the bit stream section where an error occurs in the bit stream of the reconstructed video frame based on the generated error signal. Error detection in the stream and error concealment and recovery using the same can be effectively performed.

Description

Error detection method in MPEG-2 video bit stream

1 is a schematic block diagram of a decoding system suitable for applying an error detection method in an MPEG-2 video bit stream according to the present invention.

2 illustrates a process of detecting an error occurring in an image bit stream using quantization scale information included in a macroblock according to an exemplary embodiment of the present invention, and performing error concealment and recovery based on the error detection result. One flowchart.

* Explanation of symbols for main parts of the drawings

100: system multiplexed decoding block 200: video decoding block

300: error detection block 400: error concealment and recovery block

500: frame memory

The present invention relates to a technique for detecting an error in reconstruction of a video signal encoded by compression based on an MPEG-2 algorithm capable of realizing high-definition television (HDTV). More particularly, the present invention relates to an original signal through a receiver video decoding system. The present invention relates to an error detection method suitable for detecting an error that may occur in a bit stream of an MPEG-2 image frame reconstructed with a.

As is well known in the art, the transmission of discrete video signals can maintain better image quality than analog signals. When a video signal composed of a series of image frames is represented in digital form, a considerable amount of transmission data is generated, particularly in the case of a high-definition television (HDTV). However, since the usable frequency range of the conventional transmission channel is limited, in order to transmit a large amount of digital data, it is necessary to compress the data to be transmitted and reduce its transmission amount.

Therefore, when transmitting a video signal, the transmitting system compresses and compresses the video signal by using the spatial and temporal correlation of the video signal to reduce the amount of data. To the decryption system.

As is well known in the art, MPEG-1, which only handles progressive scanning images, and whose transmission rate is approximately 1.5 Mbps, is mainly suitable for storage media. On the other hand, MPEG-2 is seen as an extension of H.261, which is mainly used for communication media and MPEG-1, which is suitable for storage media, but is widely used in a wide range of media fields such as communication, computers, consumer electronics, and broadcasting. It is an algorithm that can be applied and exchanged information. Unlike MPEG-1, MPEG-2 has the advantage of being able to handle progressive-scanning and interlaced images. . The present invention relates to an error detection technique in a bit stream occurring under such an MPEG-2 algorithm.

On the other hand, as the various compression techniques mainly used for encoding video signals in the MPEG-1 and MPEG-2 algorithms described above, a hybrid encoding technique combining a stochastic encoding technique and a temporal and spatial compression technique is known to be the most efficient. .

Most of the hybrid coding techniques described above, especially the hybrid coding technique used in the MPEG-2 algorithm, include motion compensated DPCM (differential pulse code modulation), two-dimensional discrete cosine transform (DCT), quantization of DCT coefficients, and VLC. (Variable encoding), etc. are used. Here, the motion compensation DPCM determines a motion of the object between the current frame and the previous frame, and predicts the current frame according to the motion of the object to generate a differential signal representing the difference between the current frame and the prediction value. These methods are, for example, Staffan Ericsson's Fixed and Adaptive Predictors for Hybrid Predictive / Transform Coding, IEEE Transactions on Communication, COM-33, NO.12 (Dec. 1985, December), or A motion Compensated Interframe from Ninomiy and Ohtsuka. Coding Scheme for Television Pictures, IEEE Transactions on Communication, COM-30, NO.1 (January, 1982).

Therefore, when transmitting a video signal, the transmitting side compresses and encodes the image signal in the buffer of the output side in order by considering the spatial and temporal correlation of the image signal in block units or pixel units through the above-described encoding technique. The video data, i.e., the bit stream, will be transmitted to the decoding system on the receiving side through the transmission channel at a desired bit rate in response to the channel request.

Next, in the decoding system on the receiving side, the bit stream of the compressed coded video received through the transport channel is converted into the original signal before the coded coded video data using a technique such as IVLC, inverse quantization, or IDCT. The recovered video bit stream may be recovered and concealed such as a transmission error that may occur during transmission, and then provided to the display side for display through a monitor. Such error recovery and concealment techniques are not actively provided by the MPEG-1 algorithm. However, the MPEG-2 algorithm is more aggressive in consideration of the fact that bit strings are transmitted in an ATM (Asynchronous Transfer Mode) transmission network. Recovery and concealment techniques are provided.

On the other hand, in the image encoding method using the MPEG-2 algorithm, even if only one bit is damaged in the bit stream of the image frame, the image quality degradation in the substantially reconstructed image can be greatly affected.

Therefore, detecting an error part on the bit stream of an image frame is a very important process for applying an effective error recovery and concealment technique.

Accordingly, the present invention devised in view of the above point, and can effectively detect an error in a reconstructed MPEG-2 video bit stream using quantum scale information included in macro block data on a bit stream of a video frame. An object of the present invention is to provide an error detection method in a recovered bit stream.

In order to achieve the above object, before the present invention is encoded using a decoding means including the techniques of IVLC, inverse quantization and IDCT for the video bit stream of the encoded MPEG-2 video frame separated by the system multiplexed decoding block A method for detecting an error signal generation position on a bit stream for recovering an original signal of the signal stream and concealing and recovering an error generated when an error occurs on the bit stream of the restored video frame through error concealment and recovery means, Inputting a bit stream of the reconstructed video frame to perform a codeword search for a plurality of pieces of information included in a macro block, and searching for a codeword for quantization scale information of a predetermined bit; Detecting an information value of a codeword for the retrieved quantization scale information; If it is determined in the information value detecting step that the detected quantization scale information value is an integer value other than zero, the error signal is not generated, and if it is determined that the detected quantization scale information value is zero value. Generating the error signal and providing the error signal to the error concealment and recovery means; And performing error concealment and recovery in a bit stream section in which an error occurs among the bit streams of the restored video frame based on the generated error signal. do.

Other objects and various advantages of the present invention will become more apparent from the preferred embodiments of the present invention described below with reference to the accompanying drawings by those skilled in the art.

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

As is well known in the art, the video data hierarchy of MPEG-2 is composed of a sequence layer, a group of picture (GOP) layer, a picture layer, a slice layer, a macro block layer and a block layer.

In the above hierarchical structure, the sequence layer is composed of one or a plurality of GOPs in which the sequence header information and the video rate are the same, and the GOP layer is encoded by the intra coding mode without taking a difference (error signal). A picture includes an intra coded picture, a forward predictive coded picture (P), and a bidirectionally predicted coded picture (B picture).

In addition, the picture layer is an image frame constituting one screen, and is composed of one or a plurality of slices, and the slice layer is composed of one or a plurality of macro blocks subsequent to the scanning line of the image, and the macro block layer is composed of four luminance blocks. And a macro block having two color difference blocks and determining which to use as prediction data which is reference image data which takes a difference (error signal). Substantially, such a macro block layer is intended to be pursued by the present invention. It is used for error detection in the video bit stream.

More specifically, the quantization scale information value included in each macro block in the macro block layer is quantization step size information of the macro block, and has a fixed length of 5 bits, which may have at least zero value, that is, zero data value. There is no. Detailed data of this macro block layer will be described later in detail.

Finally, the block layer consists of 8x8 lines / pixels separated by luminance or chrominance, which blocks are used as units of DCT.

Meanwhile, according to the present invention, a macroblock layer to be used for detecting an error on a bit stream of an image frame has the following data information.

1) MBE (macroblock escape): Escape code information of the macro block address, which is used only when the MBAI is 33 or more, and has a length of 11 bits.

2) MBAI (macroblock address increment): VLC indicating the number of previously skipped macro blocks + 1, information indicating the number of macro blocks + 1 from the left end of the image, having a length of 1-11 bits, In case of intra frame, it is defined as 1

3) MBT (macroblock type): VLC coding information indicating a coding mode of a macroblock, having a length of 1-9 bits, and having an intraframe having two types (1 or 1) defined according to quantization scale.

4) Spatial temporal weight code (STWC): Information indicating the upsampling method of the lower layer screen in temporal scalability, and having 2 bits in length.

5) FrMT (frame motion type): Predictive type information for a frame, MBT uses prediction, and exists only when frame predictive frame DCT (FPFD) is zero (0) in a frame structure. Has the length of

6) FiMT (field motion type): As a prediction type information for a field, MBT uses prediction and exists in the case of conditions other than FrMT and has a length of 2 bits.

7) DT (DCT type): Information indicating whether the DCT is in the frame mode or the field mode. The information is added only when the DCT data exists and has a length of 1 bit.

8) QSC (quantizer scale code): It is quantization step size information of a macro block, and it has a fixed length of 5 bits, not at least 0.

9) MVFS (motion vertical field select): In case of field prediction, information indicating which field of the reference image is predicted and has a length of 1 bit.

10) MHC (motion horizontal code): Information indicating that the difference between the horizontal component of the motion vector for the macroblock and the previous motion vector is encoded by VLC and has a length of 1-11 bits.

11) MHr (motion horizontal r): Information used with MHC and has a length of 1-8 bits.

12) DH (DMV horizontal): In the case of dual prime prediction, this is VLC information indicating a horizontal difference vector and has a length of 1 to 2 bits.

13) MVC (motion vertical code): Information indicating that the difference between the vertical component of the motion vector for the macroblock and the previous motion vector is encoded by VLC, and has a length of 1-11 bits.

14) MVr (motion vertical r): Information used with MVC and has a length of 1-8 bits.

15) DV (DMV vertical): In the case of dual prime prediction, it is VLC information indicating a vertical difference vector and has a length of 1-2 bits.

16) CBP (coded block pattern) 420: Indicates CBP information when the color difference format is 4: 2: 0, and has a length of 3-9 bits.

17) CBP1 (coded block pattern 1): Information added to CBP 420 when the color difference format is 4: 2: 2, and has a length of 2 bits.

18) CBP 2 (clded block pattern 2): Information added to CBP 420 when the color difference format is 4: 4: 4, and has a length of 6 bits.

Next, a process of detecting an error that can occur on the bit stream of an image frame in the MPEG-2 algorithm using the specific information of the macroblock layer having the above-described data information will be described.

1 is a schematic block diagram of a decoding system suitable for applying an error detection method in a bit stream of an MPEG-2 video frame according to the present invention. As shown in the figure, a typical decoding system includes a system multiplexed decoding block 100, an image decoding block 200, an error detection block 300, an error concealment and recovery block 400, and a frame memory 500. do.

In FIG. 1, the system decoding block 100 performs a function of synchronizing and separating information such as compressed coded video, voice, and text provided from a coding system of a transmitting side (not shown) through a transport channel. In this case, the separated information such as voice and text is provided to the audio and text information decoding block, not shown, and the image information in the form of a bit stream is provided to the video decoding block 200 through the line L11.

In addition, the video decoding block 200 restores a video bit stream that has been compression-coded according to a specification determined in the MPEG-2 standard to an original signal before being encoded. As is well known in the present invention, a video bit stream for an original signal is generated by reconstructing a compressed coded video bit stream using techniques such as IVLC, inverse quantization, IDCT, and motion compensation prediction. Is provided to the error detection block 300 via line L13 and simultaneously to the error concealment and recovery block 400 via line L15.

On the other hand, the error detection block 300 detects an error in the image bit stream of the reconstructed intra frame or inter frame on the line L11, and in particular, the quantization scale information included in the macro block layer on the bit stream of the image frame according to the present invention. The error position is detected in the reconstructed video bit stream using the value, and the detected error position information is an error concealment and recovery block through line L17 as a control signal for error concealment or recovery in the generated bit stream. Provided at 400.

As mentioned above, in the case of an image frame, the quantization scale information value in the macro block represents the quantization step size information of the macro block, which has a fixed length of 5 bits according to the MPEG-2 Recommendation. It has a value of zero, that is, a value other than zero.

Accordingly, in the error detection block 300, an MBE information value (not required for an image frame) having an 11-bit data value and MBA information having a data value of 1-11 bits is detected on a bit stream of an image frame. Value (must be defined as 1 for video frame), MBT information value of 1-9 bits, STWC information value of 2 bits, FrMT information value of 2 bits, FiMT information value of 2 bits, DT information value of 1 bit ( This value is only present when DCT data is present.) Checks the QSC (quantization scale code) information value following it, and if it is determined to have a value of 0, it indicates that an error has occurred in the corresponding macroblock on the reconstructed picture frame bit stream. Generate an error signal (e.g., a logic signal of 1 or 0) and provide it to the error concealment and recovery block 400 described below. As a result, in the error concealment and recovery block 400, the slice data of the reconstructed image frame including the currently input macro block (macro block in which the error has occurred) is stored in the frame memory 500. Errors will be concealed through techniques such as replacing the corresponding slice data on the spatial axis of the frame. At this time, the error concealment and recovery block 400 does not perform error concealment on the data of only the corresponding macroblock in which an error occurs, but performs error concealment on the entire corresponding slice data having the macroblock in which an error occurs. This is because it is not possible to know in which position of the slice data the macroblock in which the current error occurrence is detected is present. Therefore, error concealment and recovery are performed on the entire slice data including the macro block in which the occurrence of an error is detected.

Next, according to the present invention, the error detection block 300 detects an error in the bit stream of an image frame using a quantized scale information value having a fixed length included in the macro block, and based on the detection result, error concealment is performed. A detailed description will be made with reference to the flowchart of FIG.

First, in the error detection block 300, if a reconstructed video bit stream on the line L13 provided from the above-described image decoding block 200, that is, a bit stream for the reconstructed current frame is input (step 210), the current input is received. A codeword search for the bitword of the bit stream of the frame, for example, macroblock escape, macroblock address increment, macroblock type, DCT type, and quantization scale, is performed (step 220).

Then, in step 230, the codeword retrieved in step 220 is checked and if the currently retrieved codeword is not a codeword for quantization scale (QSC) information to be searched for error detection according to the present invention, the quantization scale If the codeword retrieval step 220 is repeated until the codeword for the information is retrieved, and it is determined that the codeword currently retrieved is the codeword for the quantization scale (QSC) information to be retrieved, the process is performed. Proceeding to step 240 proceeds.

Next, in step 240, it is checked whether the information value of the codeword for the currently retrieved quantization scale information is zero. Checking whether the information value of the codeword for the quantization scale information is 0 in this step 240 is because the quantization scale information value cannot have a value of at least zero. That is, the quantization scale information value of 0 means that an error has occurred on the bit stream. Therefore, the present invention checks whether or not an error has occurred in the corresponding macroblock by searching for the quantization scale information value.

If it is determined in step 240 that the quantization scale information value is something other than 0, the error detection block 300 generates an error plug value 0 (ie, a logic signal) (step 250). Error concealment and recovery block 400 to be provided. Accordingly, the error concealment and recovery 400 will provide the reconstructed image bit stream provided from the image decoding block 200 through the line L15 to the display side, which is not shown in the next step, without applying the error concealment and recovery technique. .

On the other hand, if it is determined in step 240 that the quantization scale information value is 0, the error detection block 300 generates an error plug value 1 (step 260) and an error of the logic signal 1 value generated in this way. The plug signal is provided to the error concealment and recovery block 400.

As a result, when the error concealment and recovery block 400 receives an error signal from the error detection block 300, the error concealment and recovery block 400 receives an error on the bit stream of the reconstructed image frame currently input from the image decoding block 200 through the line L15. Error concealment and recovery are performed on the slice data including the macroblock in which occurrence has been confirmed (step 270). For example, the error concealment and recovery block 400 replaces the corresponding slice data of the restored image frame that is currently input with the corresponding slice data in the immediately preceding image frame stored in the frame memory 500. The technique will conceal errors.

At this time, instead of performing error concealment on the data of only the corresponding macro block in which an error occurs, the reason for replacing the corresponding slice data of the previous frame with the corresponding slice data of the previous frame to conceal the error for all the corresponding slice data having the macro block in which the error occurs is currently This is because it is not possible to know in which position of the slice data the macroblock in which an error is detected is present. Therefore, error concealment and recovery are performed on the entire slice data including the macro block in which the occurrence of an error is detected. As described above, according to the present invention, when an error occurs in a bit stream of an image frame image encoded by the algorithm of MPEG-2 and then reconstructed at a receiver, an image frame is provided to conceal or recover the generated error. By using the quantization scale information in the macro block included in the bit stream of D, it is possible to effectively perform error detection and error concealment and recovery in the MPEG-2 video bit stream accompanied by image quality degradation in the reproduced video.

Claims (7)

An image bit stream of an encoded MPEG-2 image frame separated through a system multiplexing decoding block is restored to an original signal before being encoded using decoding means including techniques of IVLC, inverse quantization, and IDCT, and error concealment and A method of detecting an error signal occurrence position on a bit stream to conceal and recover an error generated when an error occurs on a bit stream of the restored video frame by means of recovery means, the method comprising: Retrieving a codeword for a predetermined bit of quantization scale information by continuously performing codeword search on a plurality of pieces of information included in a macro block by input; Detecting an information value of a codeword for the retrieved quantization scale information; If it is determined in the information value detecting step that the detected quantization scale information value is an integer value other than zero, the error signal is not generated, and if it is determined that the detected quantization scale information value is zero value. Generating the error signal and providing the error signal to the error concealment and recovery means; And performing error concealment and recovery in the bit stream section in which an error occurs among the bit streams of the reconstructed video frame based on the generated error signal. The error detection method of claim 1, wherein the quantization scale information value has a fixed length of 5 bits. The error detection method of claim 2, wherein the error signal generated based on the retrieved quantization scale information value is a plug signal having a logic value of 0 or 1. The method of claim 1, 2 or 3, wherein error detection in the reconstructed video bit stream is performed in units of macro blocks. 5. The method of claim 4, wherein error concealment and recovery in the reconstructed image bit stream are performed in units of one slice data of the currently input image frame including a macro block in which the error occurrence is detected. Error detection method in MPEG-2 video bit stream. The method according to claim 5, wherein the error concealment and recovery of the slice data in the current image frame in which the error occurrence is detected is a space in a previous image frame immediately adjacent on the time axis prestored in the error concealment and recovery block. An error detection method in an MPEG-2 video bit stream characterized by performing using corresponding slice data on an axis. 7. The MPEG-2 video bit stream of claim 6, wherein the slice data of which an error is detected in the current video frame is replaced with corresponding slice data in the previous video frame for error concealment and recovery. Error detection method.