KR100620695B1 - Method of transmitting moving picture data as a video packet - Google Patents

Abstract

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to digital interlaced video encoding and decoding, and more particularly, to a video packet unit transmission method of video data for minimizing transmission errors when transmitting a compressed video. A method of transmitting moving picture data in video packet units comprises: (a) determining whether a video packet to be coded is the first packet of a given picture; And (b) if it is determined in step (a) that the video packet is not the first packet of the given picture, transmitting a video packet header including overhead information, wherein the overhead information is the video. It is characterized by indicating whether or not interlaced decoding should be performed to decode the packet. Accordingly, even when a video packet is transmitted through an error-prone channel, the receiver can smoothly reproduce the video.

Description

Method of transmitting moving picture data as a video packet

1 is a flow chart of a time axis transmission of a general progressive video and interlaced video;

2 is an example of a video encoding type (I frame, P frame, B frame) and a prediction direction.

3 is a configuration diagram of an MPEG-4 video encoder primarily determined by the present International Standardization Organization;

4 is an example illustrating TMB, OMB and BMB;

5 is a configuration diagram of a frame MB (a) and a field MB (b) to be encoded in units of macroblocks (MB);

6 is an exemplary diagram divided into four sub-blocks within one MB;

7 is an exemplary view of a direction of referring to a previous field when performing field prediction;

8 is a structural diagram of an error resilient video packet using a resynchronization signal;

9 is a structural diagram of an error resilient video packet using data partitioning;

10 is a flowchart of encoding and transmission in data division encoding of a conventional progressive scan interpicture;

11 is a flowchart illustrating the decoding process of data division decoding of a conventional progressive scan interpicture;

12 is a flowchart showing an encoding sequence and a transmission sequence during data division encoding of an interlaced scan interpicture according to the present invention;

Fig. 13 is a flowchart showing a decoding procedure in data division decoding of interlaced scan interpicture according to the present invention.

<Explanation of symbols for the main parts of the drawings>

12: shape information encoder

13: motion estimation unit

14: motion compensation unit

15: previous object area detection unit

18: internal encoding unit

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to digital interlaced video encoding and decoding, and more particularly, to a video packet unit transmission method of video data for minimizing transmission errors when transmitting a compressed video.

In general, digital video can be divided into progressive video and interlaced video, depending on how the signal is constructed. In progressive scan video, each frame is constructed one line at a time. On the other hand, in interlaced video, each field is composed of two fields arranged one by one and then two fields are inserted one by one. Therefore, the height of each field (the number of lines) is half the height of the frame. An example illustrating this is shown in FIG. FIG. 1 (a) shows a progressive scan frame, and FIG. 1 (b) shows two fields (top field and bottom field) and interlaced scan frame. The parent and child fields are arranged one after the other, with a solid arrow for the parent field and a dashed arrow for the child field, each one representing a line, and a line between each configured field (a solid arrow). And dotted arrows) to make up the interlaced frame.

As shown in Fig. 1B, there is a time difference between the upper field and the lower field. In the case of Fig. 1B, the upper field is temporally advanced. In some cases, subfields may be temporally advanced. Due to the time difference between the two fields, the signal characteristics between adjacent lines in the interlaced frame may be different. In other words, the signal characteristic of the i th row of the interlaced frame may be different from that of the i + 1 th or i-1 th row. On the other hand, the signal characteristics of the i + 2th line or the i-1th line of the same field created at the same time may be relatively large. This characteristic becomes more prominent as the movement inside the video screen increases. Therefore, if an image coding apparatus developed according to a progressive scan video characteristic, for example, motion estimation and motion compensation, discrete cosine transform (DCT), is used for interlaced video encoding, The efficiency is reduced.

Field-based motion estimation and motion compensation, adaptive frame / field DCT, and the like have been studied to solve such a reduction in coding efficiency. In the present invention, the encoding method developed based on the interlaced video is collectively called the interlaced scanning encoding method, and the encoding methods developed based on the progressive scan video are collectively called the progressive scanning encoding method.

The interlaced scanning coding technologies are included in the MPEG-2 video coding standard, which was established for the application of digital TV by ISO / IEC JTC1 / SC29 / WG11, the world standardization organization, and are widely used in actual application products.

Meanwhile, for random access and high efficiency image compression, a method of defining and using three types of digital video frames such as an I frame, a P frame, and a B frame is widely used in a digital video encoder. In particular, the P frame and the B frame can perform a high compression ratio by performing motion compensation prediction. Here, the I frame, the P frame, the B frame, and the like are collectively referred to as a "coding type". Looking at the characteristics of the three frames in more detail as follows.

First, an I code (Intra coded frame) does not refer to other frames in encoding.

Second, since a P frame (Predictive coded frame) uses and encodes a motion compensation prediction with reference to a previous I frame or a previous P frame, a high compression rate can be encoded. The redundancy of color information is large on the time axis between successive frames of video. Therefore, rather than encoding without reference like an I frame, it estimates motion estimation between a previous I frame or a previous P frame and a current P frame, performs motion compensation prediction, and predicts error and movement information. The method of coding transmission is advantageous in terms of compression ratio.

Third, the B-frame (Bidirectionally-predictive coded frame) is the frame having the highest compression rate, and performs prediction by referring not only to the previous frame but also to the next frame. The B frame uses motion compensated prediction like the P frame. However, the compression ratio is the highest because two reference images are used and the better prediction performance is selected among them. On the other hand, this frame does not become a reference frame for another frame. Therefore, the previous frame is the previous I frame or P frame closest in display order with the current B frame, and the subsequent frame is the next I frame or P frame closest in display order with the current B frame.

2 is an example of digital video including an I frame, a P frame, and a B frame, and is shown in order of display. In FIG. 2, the upper arrow indicates the reference direction in the P frame, and the lower arrow indicates the reference direction in the B frame. In this example, the interval of I frames, the number of consecutive B frames, and the like may vary depending on the application or the equipment.

When compressing and encoding digital video, the display order and the encoding order are different from each other if I, P, and B frames are used. Taking FIG. 2 as an example, the display order is I1, B1, P1, B2, P2, B3, P3, B4, I2, whereas the encoding order is I1, P1, B1, P2, B2, P3, B3, I2 and B4. Therefore, in order to display the second frame B1 in the display order at the receiving end, it is possible after decoding the third frame P1. Therefore, applications that require low delay, such as videophones, use only I and P frames without B frames.

The example of FIG. 2 applies to both progressive scan video and interlaced scan video. However, in the case of interlaced scan video, the two fields are sandwiched between the two frames as shown in FIG. Except for this, the configuration of the video, the prediction direction, and the like are applied to both videos.

In digital video compression encoding, motion estimation, compensation, and color information encoding are performed in units of a set of pixels having a predetermined size, which is called a macro-block (MB). The most widely used MB size is 16 pixels / row x 16 lines.

MB coding methods are roughly divided into two types, intra coding and inter coding. Intra-image encoding encodes color signals owned by MB, while inter-image encoding predicts a color signal from a reference image and then encodes the error signal. Therefore, all MBs in the I frame perform intra-image encoding. On the other hand, in P frames, an encoding method that is advantageous in terms of encoding efficiency is selected. Therefore, there may be MB for intra-picture encoding in P frames.

In summary, intra-image encoding and inter-image encoding are determined depending on whether an object to be encoded is an original color signal or an error signal. In the I frame, all MBs perform only intra-image encoding, whereas in a P frame, intra-image encoded MB and inter-image encoded MB may exist simultaneously.

On the other hand, instead of encoding the entire image (frame-by-frame encoding), a video coding technique in which a frame is divided into arbitrary regions based on the information contained in the image, and encoded (object-based encoding) for each divided region. Is actively underway recently. This object-based coding technique enables users to encode and transmit only the objects they want, and this advantage will play an important role in the multimedia industry in the future.

In recent years, the world standardization organization ISO / IEC JTC1 / SC29 WG11 has been working on moving picture expert group group phase-4 (MPEG-4) standardization, which is characterized by the encoding of object units. In the video compression encoding scheme including encoding in object units, including MPEG-4, encoding and transmission of shape information representing an object region in a frame is required.

The biggest difference between the existing standards of video, for example, MPEG-1, MPEG-2, ISO / IEC JTC1 / SC29, H.261, H.263 of ITU-T, is in the shape information encoder. Able to know.

Fig. 3 shows a schematic diagram of a video encoder of MPEG-4, which is primarily determined by the present International Standards Organization.

When the VOP for each object image formed by the object region forming unit 11 is input to the MOTION ESTIMATION 13, the motion estimation unit 13 performs a macroblock unit from the applied VOP. We estimate the motion of.

In addition, the motion information estimated by the motion estimator 13 is input to a motion compensation unit 14 to compensate for the motion. The VOP whose motion is compensated by the motion compensator 14 is input to the subtractor 16 together with the VOP formed by the VOP forming unit 11 to detect a difference value, and the difference value detected by the subtractor 16 is Internal information of the object input to the object internal encoding unit 18 in units of sub blocks of the macro block.

Is encoded.

On the other hand, the VOP whose motion is compensated by the motion compensator 14 and the internal information of the object encoded by the object internal encoder 18 are added to the adder 17, and the output signal of the adder 17 is transferred. The previous VOP, which is input to the PREVIOUS RECONSTRUCTED VOP 15 and is the VOP of the image immediately before the current image, is detected.

In addition, the previous VOP detected by the previous object region detector 15 is input to the motion estimator 13 and the motion compensator 14 and used for motion estimation and motion compensation.

The VOP formed by the object region forming unit 11 is input to the shape coding unit 12 and the shape information is encoded.

Here, the output signal of the shape information encoder 12 varies depending on the field in which the VOP encoder is applied, and as shown by the dotted line, the output signal of the shape information encoder 12 is estimated by the motion estimation unit 13. ), And may be input to the motion compensator 14 and the object internal encoder 18 to be used for encoding motion estimation, motion compensation, and internal information of the object.

In addition, motion information estimated by the motion estimation unit 13, object internal information encoded by the object internal encoding unit 18, and shape information encoded by the shape information encoding unit 12 are applied to the multiplexer 19. After being multiplexed and buffered in the buffer 20, a bit stream is transmitted to a multiplexer, which is not shown in the figure, but multiplexes the outputs of a plurality of encoders and transmits them to the decoder side.

In the above MPEG-4, a coding unit having an arbitrary shape is referred to as a video object plane (VOP). This corresponds to a frame in the existing standard, and has a VOP type such as I VOP, P VOP, and B VOP as the frame type. This VOP type differs from a rectangular frame in that it has arbitrary shape information, and other characteristics (eg, movement compensation prediction directions) coincide.

Coding efficiency is reduced by applying existing techniques to encode the entire frame at the object boundary. Boundary Block Merging (BBM) and Shape-adaptive Discrete Cosine are examples of techniques studied to overcome this problem and improve coding efficiency. Transform (SA-DCT).

As described above, in digital video compression encoding, motion estimation and compensation, color information encoding, and the like are performed in units of a macroblock (MB) having a predetermined size. When encoding an object having arbitrary shape information, MB can be classified into three types with respect to shape information. First, there is MB in case it is included outside the object area. In this MB, there is no object region in the MB, and an encoding process is unnecessary. Second, it is included in the object domain, and the techniques developed in the existing frame-by-frame coding can be used directly. Finally, there is an MB in which an object region and a non-object region coexist. In order to decode such an MB in a receiver, shape information indicating an object area and color information in the object area must be encoded. Of course, in a frame for motion estimation and compensation, such as a P frame or a B frame, motion information must also be encoded and transmitted.

4 shows an example of TMB, OMB and BMB. In this case, the gray part represents an object area to be encoded, and each of the small squares shows an example of MB. In this example, there are six MBs (TMBs) contained outside the object area, twelve MBs (OMB) contained inside the object area, and twenty-two MBs (BMB) simultaneously exist inside and outside the object area.

In addition, the interlaced scan video described above exhibits different signal characteristics from the progressive scan video. In view of these characteristics, methods different from those for encoding progressive scan video have been studied for high encoding efficiency of interlaced scan video.

The interlaced scanning coding is roughly classified into two types, such as field / frame adaptive DCT and field unit motion information estimation and prediction. As a representative example of the interlaced scanning encoding method, only two of the methods described in the MPEG-4 Committee Draft (CD) are briefly described.

First, it is a field / frame adaptive DCT.

In MPEG-4, in intra-picture encoding MB or inter-picture encoding MB, the final stage of encoding is DCT (discrete cosine transform). The MB of 16x16 is divided into four subblocks having an size of 8x8, and the DCT signal of each subblock. In the case of the intra-image coded MB, the original color signal is DCTed in the inter-image coded MB. The higher the similarity between the signals of neighboring pixels, the higher the efficiency of the DCT.

In the case of interlaced scanning video, since signals of the same field are generated at the same time, there is a high probability that the similarity between signals on the same field is higher than that of the immediately neighboring pixel. This feature is the most distinguishing feature from progressive scanning video, especially when a large moving object is present in the image. In this case, it is advantageous to perform DCT between pixels of the same field. Therefore, in the case of interlaced scan video, both DCT (field DCT) between pixels of the same field and DCT (frame DCT) without considering the field are allowed as before, and a method having high compression efficiency is selected and encoded between the two DCTs. A flag indicating the selected DCT method is transmitted to the receiver. This is called field / frame adaptive DCT.

The field / frame adaptive DCT is described in detail as follows. First, we compare the similarity between neighboring lines with those between lines. If the similarity between neighboring lines is large, the frame DCT is performed. If the similarity is small, the field DCT is performed. The judgment of the similarity comparison is as follows.

In this equation, i and j denote positions of the vertical and horizontal axes in MB, respectively, and P _{i, j} denote color values of pixels at positions i and j. If the left side is larger than the right side as shown in the above relation, it means that the similarity between pixels of the same field is high. In this case, as illustrated in FIG. 5, a shuffling step ((a)-> (b)) is performed to reconfigure the upper field and the lower field to be adjacent to each other by changing the position of the line. After that, it is divided into 4 subblocks and DCT is performed. If the right side is greater than or equal to the left side, it means that the neighboring pixels have high similarity. Therefore, DCT is performed on four subblocks without performing the shuffling step. Performing DCT in FIG. 5 (a) is referred to as field DCT in performing frame DCT, and performing DCT in (b).

Information (dct_type) about the frame DCT and the field DCT is transmitted to the receiver by using 1 bit.

The decoder at the receiving end decodes the dct_type from the received bit stream and decodes the DCT transform coefficients of four subblocks to perform inverse DCT. The dct_type is then determined to determine whether it is a field DCT or a frame DCT. If the field DCT, the obtained MB is subjected to reverse shuffling ((b)-> (a)) as shown in FIG. 5 to obtain the final MB.

Second, movement information estimation in units of fields in a P frame.

The most common method of estimating motion information is to estimate motion information in a reference picture in MB units, and perform motion compensation prediction from a reproduction signal of the reference picture using this information. At this time, it is assumed that all pixels in one MB have the same movement information. Therefore, it is necessary to transmit one piece of movement information per MB to the receiver.

This assumption can be applied in the case of progressive scanning video. However, in the case of interlaced scan video, as is well known, the pixels in the same field in one MB have the same movement information, rather than the assumption that all pixels in one MB have the same movement information because the image generation time is different in one field unit. The assumption that we have is more accurate. Therefore, in the case of interlaced scan video, estimating motion information by one field enables more accurate motion information estimation, and coding gain can be obtained by reducing motion compensation prediction error. However, two pieces of mobile information must be transmitted for each MB. Therefore, in general, a better method is selected by comparing the movement information estimation in units of frames and the movement information estimation in units of fields. This selection information (field_prediction) is transmitted.

Note that as shown in FIG. 6, movement compensation of the current upper field is possible from both the upper field and the lower field of the previous frame. Therefore, information indicating the prediction direction (prediction_direction) must be transmitted. The same applies to the subfields of the current frame.

In summary, five pieces of information to be transmitted to a receiver per MB to estimate and predict field-by-field movement information are as follows.

1) Field_prediction

2) prediction_direction of the parent field

3) prediction_direction of the subfield

4) Move information of upper field

5) subfield movement information

The decoder of the receiver decodes the field_prediction information from the received bit string. In the case of performing the field-based motion compensation prediction, the 2) to 5) information may be decoded from the bit string, and the decoding may be performed according to the corresponding information.

On the other hand, when a bit string generated by compression-coding a digital video signal is transmitted through a channel in which errors may occur, for example, a mobile communication transmission error may occur, and the reception error is received by the receiving end. There is a risk that the quality of reproduction when decoding a bit stream may be greatly impaired. Researches have been actively conducted to minimize problems caused by transmission errors, and can be largely classified into error detection, error localization, and data recovery. These studies are collectively called error resilience technology, and the bit string structure reflecting these techniques is called error resilient bit string.

Hereinafter, the error tolerance recovery techniques described in the CD of the MPEG-4 encoding method for performing VOP-based encoding will be briefly described.

An error recovery technique is a resynchronization method.

The resynchronization signal enables resynchronization between the decoder and the bit string after the error is detected. If an error occurs on the bit string, the bits between the immediately preceding resynchronization point and the immediately following resynchronization point are generally ignored. If the resynchronization method can effectively localize or minimize the amount of data ignored by the decoder, it can improve the performance of other techniques related to error resilience (data recovery or error concealment).

The resynchronization method of MPEG-4 is similar to the GOB (Group of Blocks) structure used in H.261 or H.263, which is the standardization method of ITU-T video encoding. In these standards, GOB consists of one or several MB rows. When a new GOB is started in the encoding process, the GOB header is placed on the bit string. This header contains a GOB start code, which allows the decoder to know the location of this GOB. In addition to this start code, the GOB header stores the information needed to restart the decryption process. The GOB approach is spatial resynchronization. In other words, when a specific MB position is reached during encoding, a resynchronization signal is inserted into the bit string. The problem inherent in this approach is that the resynchronization signal is not located at regular intervals on the bit stream during the encoding of the variable bit-rate.

The technique adopted in MPEG-4 inserts a resynchronization signal periodically into the bit stream to avoid this problem. That is, the configuration of the video packet (VP) of the resynchronization signals is defined based on the number of bits included in the VP instead of the MB number. If the number of bits included in the current VP exceeds the given threshold, a new VP is started at the beginning of the next MB.

8 shows an example of VP. The resynchronization signal identifies the beginning of a new VP. For this purpose, this signal must be distinguished from all possible VLC codes and various start codes. Header information is also located at the beginning of the VP. This header is essential to restart the decoding process and includes information such as the location of the first MB of the corresponding VP (macroblock number) and the quantization parameter (quant_scale).

In addition, the macroblock number enables spatial synchronization, and the quant_scale enables decoding of color information after synchronization. In addition, the header extension code (HEC) is a 1-bit flag indicating whether additional information is included in the header. If HEC is 1, additional information is included in the header to enable the decoder to independently decode the VP. The additional information is as follows.

Modulo time base, VOP time increment, VOP coding type, intra dc vlc thr, VOP fcode forward, VOP fcode backward.

Here, the VOP coding type is information indicating whether the VOP to be currently encoded corresponds to one of I VOP, P VOP, and B VOP, and intra dc vlc thr represents a VLC table of DC information. The VOP fcode forward is information indicating the size of the search region when estimating movement information, and the VOP fcode backward is information indicating the size of the search region when estimating backward movement information in B VOP. Herein, the information included in the HEC is necessary information for independently decoding the VP in the decoder.

If a transmission error occurs in the VOP header information, information included in the HEC is used instead of the VOP header information. The decoder checks the semantic correspondence of the decoded information to determine whether a transmission error occurs in the VOP header.

For example, if VOP fcode forward and VOP fcode backward are decoded to a value of "0", the decoder may determine that there is an error in these information. This is because MPEG-4's fcode forward and fcode backward do not make zero. In this case, the decoder may correct VOP fcode forward and VOP fcode backward values by using the HEC of the next VP.

When using the error resilience method in MPEG-4, some compression coding methods need to be modified. For example, all prediction coding methods should limit the information used for prediction to information in the same VP to prevent error propagation. In other words, during prediction (eg AC / DC prediction, motion vector prediction), the VP boundary should be treated like a VOP boundary.

Another error recovery technique is data partitioning.

Error concealment is essential for error-resistant video codecs, and their performance depends heavily on resynchronization methods. If the resynchronization method limits the error considerably locally, the error concealment method is more effective. In low bit-rate / low delay applications, such as mobile video telephony, the current resynchronization method performs well with simple error concealment methods such as copying MBs of reference frames.

However, MPEG-4 recognizes the importance of error concealment and the need for an improved error localization method, and has developed a new technique called data partitioning. In the case of P VOP, this data partitioning separates all MB information in a video packet based on color information and movement information. Then, a second resynchronization signal is inserted between the movement information and the color information. In this case, the fact that data partitioning is used is informed to the receiver as additional information so that the decoder can know. In order to distinguish it from the resynchronization signal, the second resynchronization signal of data division is referred to as a marker signal.

9 shows a structure of a video packet in which data partitioning is used.

When the marker signal is correctly decoded, it can be said that the movement information of all MBs in the corresponding video packet is correctly decoded. If it is later detected that there is a transmission error in the transmitted and decoded color information, all the color information in the video packet is ignored and the decoded movement information before the marker signal is reproduced to reproduce all the MBs in the video packet. The motion compensation prediction is performed from the decoded reference VOP.

If movement information is important information in P VOP, DC information is important information in the case of I VOP. Therefore, all MB information in the video packet is separated around DC information and AC information. The marker signal is inserted between them. If a transmission error occurs in the AC information, the AC information is ignored and all MBs in the video packet are reproduced using only the DC information.

On the other hand, the decoder in the receiving end reproduces the original image using the transmitted bit string and a plurality of additional information. One of the decoding processes is reverse VLC coding.

The reverse variable length codes (RVLCs) are designed to simultaneously decode in a forward direction and a reverse direction. In some cases, it is possible to perform decoding in the reverse direction from a resynchronization signal immediately after an error occurrence point for a bit string in which an error exists and thus decoding in the forward direction is impossible. In this case, therefore, the number of bits ignored by the decoder by the transmission error is reduced. Currently, MPEG-4 applies RVLC only to encoding of transform coefficients.

10 is a flowchart of data segmentation encoding of a progressive scan interpicture of MPEG-4, which is currently undergoing standard drafting.

This flowchart shows the order of information transmitted in one video packet. When the number of MBs present in this packet is N _MB , first, all moving information of N _MB MBs belonging to the packet is encoded and transmitted first. And transmits the motion_maker (motion_maker), and also encodes the color information of the N _MB MB

send. However, since MPEG-4 is encoded in an area unit called a VOP having arbitrary shape information, shape information of each MB is first transmitted together with movement information, and additional information for decoding motion information is also transmitted first. The movement_maker is a signal that distinguishes the movement information from the color information and is a signal that should be distinguished from all possible codes in all encoders.

FIG. 11 is a flowchart illustrating a case of performing data division decoding on a progressive scan interpicture in MPEG-4, which is currently undergoing standard drafting.

The decoder first decodes the shape information, and then decodes additional information necessary for decoding the motion information. The motion information is then decoded with additional information necessary for the decoded motion information. This process is repeated until the next bit string is move_maker. At this time, the number of repetitions is the number of MBs present in the video packet, N _MB . After decoding the " move_maker &quot;, color information for all N _MB MBs is decoded together with additional information.

As additional information for the movement information considered in the image processing system, a flag ("not_coded") indicating that the MB does not have any coded bits at all, and one mode per MB if the movement information is a certain mode (progressive scanning). Information indicating whether the mobile information is sent and the four modes are sent). The additional information for the color information includes information indicating which sub-blocks have a color coefficient to be transmitted, a flag indicating whether the color coefficient is predictively encoded.

On the other hand, all existing techniques for error resilience described so far are developed for video of progressive scan P frame.

Therefore, a large number of errors occur when the conventional technique of encoding and decoding the progressive scan interpicture is applied to the interlaced scan interpicture.

For example, when encoding a progressive scan interpicture, it is not necessary to distinguish each field. However, when encoding an interlaced video, each field needs to be distinguished. Therefore, the encoding method of encoding a progressive scan video may be performed as it is. When applied to Interpicture, many errors occur.

Accordingly, an object of the present invention is to provide a video packet transmission method of moving picture data in which smooth encoding and decoding is performed when encoding or decoding interlaced scan video, and in particular, minimizing a transmission error when transmitting a compressed video.

In order to achieve the above object, there is provided a method of transmitting moving picture data in video packet units, comprising: (a) determining whether a video packet to be coded is the first packet of a given picture; And (b) if it is determined in step (a) that the video packet is not the first packet of the given picture, transmitting a video packet header including overhead information, wherein the overhead information is the video. It is characterized by indicating whether or not interlaced decoding should be performed to decode the packet.

Hereinafter, with reference to the accompanying drawings, preferred embodiments according to the spirit of the present invention as described above in detail as follows.

First, in the present invention, when interlaced scanning encoding is performed, interlaced information included in the video packet is included. Furthermore, additional information required should be transmitted. In case of performing interlaced decoding, when a transmission error occurs in the interlaced transmission part, it is possible to determine whether to perform interlaced decoding in the decoder by decoding the interlaced in the packet header.

For example, in the case of MPEG-4, interlaced information is included in the VOP header, and information related to the scan method (alternate_scan) and information indicating which field among the upper field / lower field are temporally first (top_field_first) are included in the interlaced encoding. Transmit as necessary additional information. Therefore, in this case, when the HEC is 1, the encoder must transmit three pieces of information including interlaced, alternate_scan, and top_field_first in the video packet header. In the decoder, if a transmission error occurs in the VOP header portion, the above three pieces of information in the video packet are decoded and used for decoding the corresponding video packet.

Next, in the data segmentation of the interpicture in the compression encoding of interlaced scan video, movement information and color information of all MBs in a predetermined video packet are transmitted separately. In addition, a marker signal is inserted between the movement information and the color information to determine whether an error occurs when the movement information is transmitted.

If an error is detected in the color information, the color information is ignored and all MBs in the video packet are reproduced by performing motion compensation prediction from the previous frame using the decoded motion information before the marker signal.

On the other hand, in the inter picture interlaced coding scheme, the motion information to be transmitted to the receiver includes a flag bit indicating whether the motion compensation prediction is in the field mode or the frame mode, and in the field motion compensation prediction mode, Movement information and information indicating a prediction direction of each field.

In the case where the intra-picture coded MB and the inter-picture coded MB exist at the same time as the frame type, there are a P frame and a B frame. The marker signal inserted between the information for the field-by-field motion compensation prediction and the color information is selected by all the variable length codes generated by the encoder, signals for various resynchronizers, and signals that are not overlapped with the start signal. Do not confuse it with other signals.

In the above, examples of the information for the field-based motion compensation prediction are mode information indicating whether to predict the field-based motion compensation, and information indicating the movement direction and the prediction direction of each field when the mode is a field unit.

In addition, various methods are available for the decoder to determine the length of a video packet (the number of MBs present in the video packet). For example, information for field-based motion compensation prediction may be decoded until the next bit string to be decoded is a marker signal, and the length of each video packet may be represented by the number of MBs and transmitted as additional information.

If a transmission error occurs in the color information, the color information is ignored and the prediction is performed from the reference frame using the information for the field-based motion compensation prediction, which is already decoded and stored, and then the reproduction MB is composed of these prediction values. .

Hereinafter, embodiments of the present invention will be described in more detail.

12 is a flowchart illustrating a process of splitting and encoding interlaced data between inter-images and inter-images using an interlaced scanning encoding method according to the present invention.

As shown in the figure, the coding order and transmission order of information in one video packet are shown. First, the shape information of the i-th MB of a predetermined video packet is encoded at the start position of the VP, and is first transmitted to the decoder side. Next, additional information (eg, information indicating an interlaced scan frame, information on a scan method, etc.) for movement information is encoded and transmitted to the decoder. A flag (field_prediction) bit indicating whether the motion compensation prediction is in the field mode or the frame mode is encoded and transmitted to the decoder. At this time, when the motion compensation prediction mode is the field motion compensation prediction mode, the motion information of each field and the information indicating the prediction direction of each field are encoded and then transmitted to the decoder. When the motion compensation prediction mode is the frame mode, the frame motion information is encoded and transmitted to the decoder. In the present invention, for example, when the field_prediction has a value of "1" indicates that the field mode. These processes are performed for all N _{MB MBs} in the video packet. A marker signal (motion_marker) for distinguishing the motion compensation prediction information and the color information is encoded and transmitted to the decoder. Finally, the color information of all N _{MB and} all MB information and additional information necessary for the encoding are transmitted to the decoder, thereby ending the intra- and inter-image encoding and transmission of the interlaced scanning interpicture. .

13 is a flowchart illustrating a decoding process when the interlaced and inter-picture data is divided and decoded by an interlaced scan interpicture according to the present invention.

As shown in the figure, first, shape information of the i-th MB in the received predetermined video packet is decoded, and then movement information of the MB is decoded with additional information. Next, a flag (field_prediction) indicating whether a field mode or a frame mode is decoded for the motion compensation prediction of the corresponding MB is decoded. In this case, as a result of decoding the flag, information necessary for the field-based motion compensation prediction is decoded in the field mode, and in contrast, frame motion information is decoded in the frame mode. This process is repeated until the next bit string indicates motion_marker. At the same time calculate the number of iterations, ie N _MB . After decoding the motion_marker that distinguishes the movement information from the color information, the original image of the MB is decoded by decoding the color information corresponding to the N _MB MBs and the additional information therefor.

Will be reproduced.

In the embodiment of the present invention, it has been described that the process of encoding or decoding the shape information, the additional information, and the movement information is repeated until the next bit string indicates the motion_maker. When transmitting, the decoding side does not have to check the motion_marker every time, and repeats the decoding process of shape information, additional information, and movement information as much as the number of MBs as in the encoder.

Since the present embodiment is applied to an encoder / decoder of MPEG-4 having arbitrary shape information, the shape information transport block has been described as encoding and decoding first in the embodiment. However, the method of encoding and decoding the interlaced interpicture according to the present invention in-picture and inter-picture may be applied to a coding method of a frame unit of a rectangle. In this case, only the process of encoding shape information in FIG. 12 and the process of decoding shape information in FIG. 13 may be eliminated.

When encoding and decoding are performed using the proposed method of the present invention as described above, a high quality interlaced video can be applied to a medium transmitting a channel having a high probability of transmission error. . Taking digital TV as an example, digital TV transmits high quality interlaced scan video using airwaves, and the airwave channel has a high probability of transmission error. When the TV decodes the bit stream of the compressed image, the encoding and decoding method of the present invention as described above has the advantage of obtaining a high quality reproduction image even in the presence of a transmission error.

As described above, according to the present invention, in the case of intra- and inter-image encoding interlaced scanning interpicture, the present invention is encoded by using a data division encoding method, and the encoded video packet of the interlaced scanning interpicture is encoded. The decoder transmits the decoder to the decoder according to the transmission order proposed by the present invention. The decoder performs the decoding according to the method proposed by the present invention, even when the video packet is transmitted through a channel in which a channel error occurs. There is an effect that can be reproduced sufficiently.

Claims (6)

In the method for transmitting video data in video packet units, (a) determining whether the video packet to be coded is the first packet of a given picture; And (b) if in step (a) it is determined that the video packet is not the first packet of the given picture, transmitting a video packet header including overhead information, The overhead information indicates whether or not interlaced decoding should be performed to decode the video packet. The method of claim 1, And the video packet comprises a plurality of macro blocks. The method of claim 2, And the picture unit interlaced overhead information comprises a parameter for interlaced decoding. The method of claim 3, The parameter comprises a top field first and an alternate vertical scan flag alternate_vertical_scan_flag. The method of claim 3, The picture unit interlaced overhead information includes coding information indicating whether the video packet is encoded as an intra picture or an inter picture. The method of claim 5, If the video packet is encoded as an intra picture, the coding information indicates an interlaced intra picture, and when the video packet is encoded as an inter picture, the coding information indicates an interlaced inter picture.

Images