KR0148146B1 - Apparatus for loss-prevention of important image data - Google Patents

Abstract

The present invention relates to a digital image data decoding apparatus. In more detail, when decoding the edited video data, a lot of overflow occurs in the buffer of the decoding side. If an overflow occurs in the buffer, all incoming video data is discarded. At this time, important video data is lost, which adversely affects image quality. To prevent this, compare the amount of data stored in the buffer with the amount of data by the image data header information, and detect the error value.The amount of data stored in the buffer is larger than the amount of data by the image data header information. In this case, the present invention relates to an apparatus for preventing loss of important video data, which searches for and removes data of bidirectional predictive encoding data having a data amount most similar to the error value in advance. Therefore, it is possible to prevent the loss of the data of the intra mode encoded screen or the forward predicted encoded screen, which is more important than the data of the bidirectional predicted encoded screen.

Description

Loss prevention device of important video data

1 is a block diagram showing an apparatus for encoding / decoding a general video signal.

FIG. 2 is a conceptual diagram illustrating continuous screen data encoded by the encoding apparatus of FIG. 1 (a) and transmitted to the decoding apparatus of FIG. 1 (b).

3 is a block diagram showing an apparatus for preventing the loss of important video data according to the present invention.

4 is an explanatory diagram for explaining a decoding side buffer (Buffer) state by the apparatus of FIG.

* Explanation of symbols for main parts of the drawings

20: decoding side buffer 30: header information detection unit (virtual buffer data calculation unit)

31: comparator 32: buffer data amount calculation unit

33: removal elimination unit 34: B screen search unit

35: B screen data amount calculation unit 36: B screen data amount removal unit

The present invention relates to an apparatus for preventing the loss of important video data, and in particular, to prevent an overflow occurring in the decoding side buffer when decoding two different bit streams. The present invention relates to a loss prevention apparatus of important video data which can prevent the loss of important video bitstreams.

Recently, in order to improve image quality, a method of encoding and transmitting a video signal to digital data and decoding and processing the transmitted digital data again has become common. However, when the video signal is encoded by digital data, the data amount is considerably large. Therefore, conversion encoding, differential pulse code modulation (DPCM), vector quantization, variable length coding, etc., are performed to remove redundancy data included in the digital video signal.

FIG. 1 (a) is a block diagram showing an apparatus for encoding a general video signal. In general, an apparatus for encoding a video signal includes means for quantizing a transform coefficient after performing a DCT transform on an N × N capable force, means for further compressing the amount of data by variable length encoding the quantized data, and variable length encoding. It consists of an encoding side buffer that stores image data and transmits it to the decoding side at a constant speed, and means for performing motion prediction and motion compensation from a reconstructed picture and a block to be encoded by inverse quantization and inverse transformation of quantized data.

In more detail, NxN-sized block image data (generally N ₁ xN ₂ blocks or N ₁ = N ₂ = N) is applied to the input terminal 10. The block data applied through the input terminal 10 is subtracted from the predetermined feedback data by the first adder A1 to calculate error data. This error data is converted into data in the frequency domain by the NxN converter 11. The quantization unit 12 converts the transform coefficients into representative values of a predetermined level through a predetermined quantization process. That is, the quantization unit 12 receives a quantization level (Q Level) signal fed back from the encoding side buffer 14, thereby quantizing the output data of the NxN converter 11.

The variable length encoder 13 compresses the image data by variable length encoding the quantized coefficients according to their statistical characteristics. Variable-length encoded video data is supplied to the encoding side buffer 14 in the form of a bit stream. The encoding side buffer 14 stores the bit stream type image data supplied from the variable length encoding unit 13 and transmits the image data to the decoding side at a constant speed. In addition, the encoding side buffer 14 supplies a quantization level (Q Level) signal to the quantization unit to adjust the amount of input data so that overflow or underflow does not occur.

On the other hand, in general, since there are many similar parts between the screen and the screen, in the case of a screen having a slight movement, the motion vector is calculated by estimating the motion, and the data is compensated using the motion vector. In order to perform such dynamic compensation, the inverse quantization unit 15 and the NxN inverse transform unit 16 dequantize the quantized coefficients quantized by the quantization unit 12 and perform NxN inverse transformation. The video data output from the NxN inverse transform unit 16 is added to the predetermined feedback data by the second adder A2 and stored in the frame memory 17 to reconstruct the screen. Then, in the tracking unit 18, block data having a pattern most similar to the NxN block data applied from the input terminal 10 is found in the frame data stored in the frame memory 17, and the motion vector MV representing the movement between two blocks is obtained. Calculate. This moving vector is transmitted to the decoding side and the compensating unit 19. The compensating unit 19 reads block data corresponding to the motion vector from the frame data of the frame memory 17 and supplies it to the first adder A1. Then, the first adder A1 calculates the difference data between the block data applied from the input terminal 10 and the block data supplied from the dynamic compensator 19. The difference data is variably encoded by NxN conversion and quantization, stored in the encoding side buffer 14, and then transmitted to the decoding side. Here, the two switches SW1 and SW2 are refresh switches for refreshing data in a frame unit or a predetermined block unit in order to prevent the encoded image from being different from the actual image. In other words, this switch determines the encoding mode of the intra mode or the inter mode.

FIG. 1 (b) is a block diagram schematically illustrating an apparatus for decoding a general video signal. As shown, a decoding side buffer, a variable long decoding unit, an inverse quantization unit, an NxN inverse transform unit, an adder and a frame memory are provided. The image data transmitted at a constant speed from the encoding side is stored in the decoding side buffer 20. The variable length decoding unit 21 extracts the image data of the screen to be constructed from the decoding side buffer 20 and decodes the data through the reverse process of variable length encoding. The decoded image data is input to the inverse quantization unit 22, and the inverse quantization unit 22 inversely quantizes the image data using the transform coefficient of the frequency domain. The NxN inverse transform unit 23 converts the transform coefficient of the frequency domain supplied from the inverse quantizer 22 into image data of the spatial domain. In addition, frame data stored in the frame memory 24 corresponding to the motion vector transmitted from the encoding tracking unit 18 on the encoding side is supplied to the adder A to generate image data of the spatial domain generated by the NxN inverse transform unit 23. Are combined and sent to the display.

FIG. 2 is a conceptual diagram illustrating continuous screen data encoded by the encoding apparatus of FIG. 1 (a) and transmitted to the decoding apparatus of FIG. 1 (b). As shown, a plurality of I screens, B screens, and P screens are arranged in a predetermined order.

The I screen is a screen in which all block data of one screen is encoded in the intra mode. That is, the I screen is a variable length coded screen through NxN conversion and quantization for each block irrespective of the other screens before and after the image data in one screen. The I screen is a screen closer to the original image than a screen obtained by encoding difference values with other screens. The coding efficiency is lower than that of the screen encoding the difference value between the screens. However, if the screen is inserted between the screens and transmitted to the decoding side, the decoding side can perform random access or high speed playback.

The P screen is the forward predictive encoding screen. That is, the P screen is a screen configured by selecting appropriate screen data in unit of block size among screen data encoding difference value from previous screen generated by motion estimation and motion compensation and screen data encoding original screen. .

Screen B is a bidirectional predictive encoding screen. That is, the screen B is a block unit of a predetermined size of the screen data encoding the difference value between the I, P screen before and after the B screen or the screens generated by motion compensation for the interpolation screen and the screen data encoding the original screen. This screen is composed by selecting appropriate screen data. Since the B screen is not used for any other screen reproduction other than the own screen reproduction, even if the data of the B screen is lost on the decoding side, the effect is less than that of the I or P screen. These I, P, and B pictures are encoded in a predetermined order and transmitted to the decoding side. In general, the order of pictures to be encoded is different from the order of pictures to be transmitted to the decoding side. In the encoder, the first fragment data is encoded into an I picture. The next screen data is encoded into a P screen. The B picture is encoded based on the screen data of the I picture and the P picture. That is, the encoding order in the encoder is in the form of IPBBPBB ... However, the transmission order of the encoded picture data is in the form of IBBPBBP ... as shown. In other words, the order in the encoder is changed.

Data of each screen such as I, B, and P is generated in the form of a bit stream in the encoder. The rate of occurrence of the bitstream generated by the encoder is not constant. Therefore, the bitstream generated by the encoder is stored in the buffer of the encoding side and then transmitted to the decoding side by the buffer at a constant speed.

The bitstream transmitted from the encoding side is stored at a constant speed in the buffer at the decoding side. The bitstreams stored in the buffer of the decoding side are extracted by a predetermined amount by the decoder to form respective screens through the decoding process. At this time, the encoder records information (virtual buffer data amount) of the data amount to be stored in the decoding side buffer at the time of decoding the corresponding picture in the front part (header information) of the bit stream of each screen. In an unedited bitstream, the amount of data that is filled in the actual buffer and the amount of data that should be filled in the buffer by the header information at each point of time coincide. However, when editing using several different bitstreams, there are discontinuities where different bitstreams are connected.

For example, there is a case where a bitstream of a picture and a bitstream of a picture of b are edited into a bitstream of a picture of c. In this case, the c picture bitstream is used as a video bitstream up to a certain portion, then the b picture bitstream is used up to a predetermined portion, and then the a video bitstream is used up to a predetermined portion. This can be used alternately. The portion where the a video bitstream and the b video bitstream are connected becomes a bitstream discontinuity. If a video bitstream is used to compose a screen until a certain point, and then suddenly b video bitstream is used, a larger amount or smaller amount of a video bitstream must be escaped in a buffer at the decoding side for the screen composition. b The image bit stream is exited. Therefore, at that point in time, the amount of data that is filled in the actual buffer does not match the amount of data that must be filled in the buffer by the header information. This increases the likelihood that an overflow or underflow will occur in the buffer, and in practice there are many cases. When overflow occurs in the decryption side buffer, the newly introduced bitstream is discarded. At this time, a bit stream of I picture or P picture data, which is not B picture data, is inevitable. If the I or P picture data is lost, the picture quality will be very poor.

An object of the present invention is to predict the possibility of overflow in the buffer on the decoding side when decoding the edited video data, thereby selectively removing the B screen data, thereby preventing the overflow of data of the I screen or the P screen. It is to provide a loss prevention device of important video data that can reduce deterioration of image quality.

As described above, an object of the present invention is to store digital video data encoded and transmitted in a decoding side buffer, and then to decode the information. The information of the first video data amount to be stored in the buffer from the transmitted video data (virtual buffer) Data amount), and means for comparing the first image data amount with the second image data amount substantially stored in the buffer to detect an error value thereof, and the second image data amount from the error value detection means. Means for selecting and removing image data having a data amount similar to the error value having the least influence on the image quality among the transmitted image data when the error amount is detected because the amount is greater than the first image data, and the error value detection According to the presence or absence of the error value detected by the means, the buffer is input to the image data and the output terminal of the image data removal means. Is accomplished by either a critical characterized in that it comprises a selection switch for connecting the image data loss protection.

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

3 is a block diagram showing an apparatus for preventing loss of important video data according to the present invention. As shown in the figure, when image data in a bitstream form is transmitted from the encoding side, the header information detector 30 detects information on the amount of data to be stored in the decoding side buffer 20 loaded in the bitstream and compares the comparator ( 31). At this time, the buffer data amount calculating unit 32 calculates the amount of data filled in the decoding side buffer 20 and supplies it to the comparator 31. Then, the comparator 31 calculates an error value between the amount of data by the header information detection unit 30 and the amount of data by the buffer data amount calculating unit 32 and supplies it to the removal determining unit 33. In addition, the comparator 31 operates the switch SW in accordance with the presence or absence of an error value.

On the other hand, when the image data in the bitstream form is transmitted from the encoding side, the B screen search unit 34 searches for the B screen data and supplies it to the B screen data quantity calculation unit 35. Then, the B screen data amount calculation unit 35 calculates only the data amount of the B screen of each screen data and supplies it to the removal existence determining unit 33. The removal existence determination unit 33 designates B screen data that can minimize the absolute value of the error value supplied from the comparator 31 among the B screen data, and sends the information to the B screen data removal unit 36. The B screen data removal unit 36 removes the designated B screen data and sends the rest to the decoding side buffer 20.

4 is an explanatory diagram for explaining a decoding side buffer state by the apparatus of FIG. In FIG. 4 (a), changes in time of different bitstreams a and b are shown on a rectangular coordinate plane having a time axis and a buffer data amount axis.

The portion where the amount of data increases with the change of time with a constant slope indicates the state where the data flows into the buffer. The reason for having a constant slope is that data is transmitted at a constant speed in the buffer on the encoding side. The vertical decrease in the amount of data in the buffer at a certain point in time iT, (i + 1) T, (i + 2) T, ... is assuming that the data to compose the screen is simultaneously drawn out of the buffer. . That is, the amount of data simultaneously exiting from iT, (i + 1) T, (i + 2) T, ... is the amount of data corresponding to one screen. For example, if the data exiting the buffer at (i + 3) T is the data of the I screen, the amount of data that decreases vertically at the (i + 3) T time is the data amount of the I screen.

FIG. 4 (b) shows the amount of data exiting the buffer, respectively. As shown in bold, the bold line is for the a bit stream in FIG. 4 (a) and the thin line is for the b bit stream in FIG. 4 (a). The data amounts of the a and b bit streams exiting the buffer at different time points iT, (i + 1) T, (i + 2) T, ... are different. For example, the difference in the amount of data in the a and b bit streams exiting the buffer at (i + 3) T is q.

FIG. 4 (c) shows the amount of change with time in the buffer of the c bit stream edited into the a and b bit streams. That is, it shows that a bitstream is used until the (i + 2) T time point, and after that, the cbitstream using the bbitstream.

As shown, up to (i + 2) T is indicated by a solid line, followed by dashed lines and solid lines, respectively. Here, the part indicated by a dotted line represents a general case not according to the present invention, and the part indicated by a solid line represents the case according to the present invention. The amount of data exiting the buffer at time iT, (i + 1) T, (i + 2) T is out of the buffer at time iT, (i + 1) T, (i + 2) T in FIG. 4 (b). It is equal to the data amount of the outgoing a bitstream. In the case of the portion indicated by the dotted line after the (i + 3) T time, the amount of data to be removed from the buffer at the (i + 3) T, (i + 4) T and (i + 5) T time is the fourth (b). This is equal to the amount of b-bit streams exiting the buffer at that time. That is, at the point (i + 3) T, only the amount of b-bit streams of FIG. 4 (b) exits from the portion to be escaped from the buffer by the amount of a-bit streams of FIG. 4 (b). Therefore, at time point (i + 3) T, as much data as q in FIG. 4 (b) remains, and at a later point in time, the data is stored in the buffer by the information recorded in front of the bitstream of the screen unit transmitted to the encoding side. The amount of data to be filled and the amount of data filled in the actual buffer are inconsistent by q in FIG. 4 (b). As a result, the possibility of overflow in the buffer increases, and when an overflow occurs, the bitstream flowing into the buffer thereafter is discarded in units of screens without distinguishing between I, P, and B screens. At this time, the error value (q) of the amount of data to be filled in the buffer at (i + 3) T and the amount of data to be filled in the buffer recorded in front of the bitstream to be introduced into the buffer immediately after (i + 3) T ), And after selecting and removing the B picture data having the amount most similar to the error value q among the bit streams to be introduced into the buffer in advance, it is as indicated by the solid line and no overflow occurs. The amount of B picture data to be removed by the vertically decreasing portion between (i + 3) T and (i + 4) T, and has a value similar to the error value q.

Even if there is a lot of editing, you can prevent the overflow from happening in the buffer by repeating the above process. Therefore, it is possible to prevent the loss of the data of the I or P screen and to reduce the deterioration of image quality due to editing.

Claims (5)

A device for decoding and storing digital image data encoded and transmitted in a decoding side buffer, comprising: detecting information (virtual buffer data amount) of a first image data amount to be stored in the buffer from the transmitted image data, Means for comparing the first image data amount with the second image data amount substantially stored in the buffer and detecting an error value thereof; When the error amount is detected because the second image data amount is larger than the first image data amount from the error value detecting means, the image data having the smallest influence on the image quality among the transmitted image data and similar to the error value is detected. Means for selecting and removing; And a selection switch unit for connecting the buffer to one of an input terminal of the image data and an output terminal of the image data removing unit according to the presence or absence of the error value detected by the error value detecting unit. Loss prevention device. 2. The apparatus of claim 1, wherein the means for detecting the error value comprises: an information detector for detecting information on the first image data amount, a buffer data amount calculation portion for calculating a second image data amount stored in the buffer; And a comparator for comparing the first image data amount and the second image data amount to detect the error value. 2. The image data removing unit of claim 1, wherein the image data removing unit comprises: an image data amount calculating unit for calculating a data amount of image data groups having the least effect on image quality, and the data of the image data groups calculated by the image data amount calculating unit. And a removal / determination unit for determining an image data group to be removed by comparing the amount and the error value detected by the detection means, and an image data removal unit for removing the image data group determined by the removal / non-determination unit. Loss prevention device of important image data characterized by. The apparatus of claim 1, wherein the image data group to be removed is bidirectional predictive encoding screen data. The apparatus of claim 1, wherein the image data group removed by the removing means is image data in units of screens.