MXPA00008745A - Method and apparatus for encoding a motion vector of a binary shape signal - Google Patents

Abstract

An apparatus encodes a video signal composed of texture information and shape information by selectively using a progressive and an interlaced coding technique. The shape information is divided into a multiplicity of BABs of M x N pixels having one of a first and a second binary values and the texture information has a plurality of macroblocks of the same number of pixels as a BAB. Specifically, the apparatus determines a motion vector of a target BAB of the shape information based on a corresponding motion vector predictor selected among candidate motion vectors including frame-based and field-based motion vectors for BABs which surround the target BAB and have been previously encoded and frame-based and field-based motion vectors for macroblocks which surround the macroblock corresponding to the target BAB and also have been previously encoded according to predetermined priority, and encodes the motion vector of the target BAB.

Description

METHOD AND APPARATUS FOR CODING A VECTOR OF MOVEMENT OF A SE IN BINARY TECHNICAL FIELD OF THE INVENTION The present invention relates to a method and an apparatus for encoding a binary signal; and, more particularly, to a method and an apparatus for encoding a signal movement vector of binary form. PREVIOUS TECHNIQUE In digitally televised systems such as video telephone, teleconferencing and high definition television systems, a large amount of digital data is required to define each video frame signal, since a video signal line in the signal of video frames comprises a sequence of digital data referred to as pixel values. Since, however, the available frequency bandwidth of a conventional transmission channel is limited, in order to transmit the large amount of digital data it is necessary to compress or reduce the volume of data through the use of various compression techniques. data, especially in the case of low bit rate video signal encoders such as those of video telephones and teleconferencing systems. One of these techniques for encoding video signals for a low bit rate coding system is the so-called coding technique with object-oriented synthesis-analysis. Where a feed video image is divided into objects and three sets of parameters to define movement, contour and pixel data of each object that are processed through different coding channels. An example of this object-oriented coding scheme is the so-called Motion Picture Experts Group (MPEG) 4 (MPEG-4), which is designed to provide an audio-visual coding standard for allow interactivity based on content, improved coding efficiency and / or universal accessibility in applications such as low bit-rate communication, interactive multimedia (eg games, interactive TV, etc.) and area surveillance. According to MPEG-4, a video feed image is divided into a plurality of video object planes (VOP'S = Video Object Plans) which corresponds to entities in a stream of bits that a user can access and manipulate. A VOP can be referred to as an object and represented by a border rectangle whose width and height can be multiples smaller than 16 pixels (a macroblock size) surrounding each object, so that the encoder can process the video image fed into it. a base VOP-by-VOP. A VOP described in MPEG-4 includes shape information and texture information consisting of luminance and chrominance data, wherein the shape information represented in binary form signals is referred to as an alpha plane. The alpha plane is divided into a plurality of alpha binary blocks, where each binary alpha block (BAB) has 16 x 16 binary pixels. Each of the binary pixels is classified either as a background pixel or an object pixel, where the background pixel located outside the object in the alpha plane is used to assign a binary pixel value, for example 0, while that the object pixel within the object is used to assign another binary pixel value, for example 255. Each of the binary pixels in the BAB can be encoded by using a conventional bitmap-based shape coding method such as a discipline of context-based arithmetic coding (CAE = Context-based Arithmetic Encoding). For example, in an intra-mode, all binary pixels of a BAB are coded by using an intra-CAE discipline to thereby generate an intra-coded BAB, where a context value for each binary pixel of the BAB in intra-domain discipline CAE is calculated by using pixel binary values of a predetermined number, for example 10, of binary pixels that surround each binary pixel in the BAB. In an intermode, on the other hand, all the binary pixels of a current BAB are coded using an inter-CAE discipline to generate an inter-coded BAB, where a context value of each binary pixel of the current BAB in the discipline inter-CAE is calculated using values of binary pixels of a predetermined number, for example 4, of binary pixels surrounding each binary pixel in the current BAB and binary values of a predetermined number, for example 5 of binary pixels within a BAB compensated in movement with border (see MPEG-4 Video Verification Model Version 7.0 International Organization for Standardization, Coding of Moving Pictures and Associated Audio Information, (Video Verification Model MPEG-4 version 7.0, International Organization for Standardization, Coding of Images in Motion and associated Audio Information) ISO / IEC JTC1 / SC29 / WG11 MPEG97 / N1642, Bristol, April 1997, pp. 28-30). Meanwhile, the coding discipline of conventional binary form, a signal so representing or characterizing a corresponding coding condition for the BAB is encoded, in order to improve coding efficiency, in this way by generating and then transmitting a coded mode signal correspondent . For example, if all binary pixels within a BAB are object pixels, instead of encoding the binary pixel values of the object pixels to generate encoded binary pixel values to be transmitted, it will be preferable to encode a mode signal, indicating that all binary pixels within the BAB are object pixels. By employing the method described above, ie by transmitting the corresponding coded mode signal as binary information for the BAB, it is possible to improve the coding efficiency. With reference to Table 1, there are 7 mode numbers for the alpha-binary information of a BAB according to a coding discipline in conventional mode, wherein a motion vector difference for form (MVD) of BAB is a difference between a motion vector for form (MV) and a motion vector forecaster for form (MVP); and the MVP is determined by using a conventional motion estimation discipline (see MPEG-4 Video Verification Model Version 7.0, International Organization for Standardization, Coding of Moving Pictures and Associated Audio Information, (MPEG-4 Video Verification Model, version 7.0). , International Organization for Standardization, Coding of Moving Images and Associated Audio Information) ISO / IEC JTC1 / SC29 / WG11 MPEG97 / N1642, Bristol, April 1997, pp. 20-23).

Table 1. Coding Modes Mode 0 represents that an MVD for a current BAB is zero and the current BAB can be represented by a more similar candidate BAB, while mode 1 indicates that an MVD for a current BAB has a different value than zero and the current BAB can be represented by a more similar BAB candidate. For the current BAB of the zero mode, only the mode signal is encoded; while the current BAB of mode 1 is represented by the mode signal and its MVD. When determining "no_update", a difference BAB is formed by a difference between each pixel of the most similar BAB and a corresponding pixel in a current BAB and an error is verified for any of the 4 x 4 sublocks of 4 x 4 pixels included in the difference BAB is less than a predetermined threshold, an error of a sub-block is, for example, a sum of absolute pixel values within the sub-block. If the error values of all sub-blocks are equal to or less than the threshold, the current BAB is declared as mode 0 or 1 depending on the value of its MVD. Similarly, if an error for any of the x 4 sub-blocks is equal to or less than the threshold when the pixels in the current BAB are all changed to 0, the current BAB is coded as an "all_0" mode, ie mode 2 If an error for any of the 4 x 4 sub-blocks is equal to or less than the threshold when the pixels in the current BAB are all changed to 255, the current BAB is coded as an "all_255" mode, ie mode 3 For the current BAB of mode 2 or 3, only the mode signal is encoded for the current BAB. When a current BAB does not belong to any of the modes 0 to 3, either "intra-CAE" or "inter-CAE" is used for the current BAB encoding, while a BAB current mode 4 is represented by the signal mode and BAB data encoded intra-CAE. A current BAB of mode 5 is represented by the mode signal and BAB data encoded inter-CAE; and a BAB current of mode 6 is represented by the mode signal, BAB data encoded inter-CAE and an MVD. Conventional binary coding methods employing the conventional coding method described above, are basically progressive coding methods, i.e., in conventional binary form coding methods, an interlaced coding technique performed by using a method of field-based movement estimation, has not been used. However, in order to improve the coding efficiency when a spatial and / or temporal correlation between frames is smaller than that of fields, the interlaced coding technique has adapted to the coding of the signal in a binary manner. An exemplary method and apparatus for adaptively encoding the signal in a binary manner by using progressive coding and interlaced coding techniques is described in a commonly owned co-pending application, U.S. patent application. Serial No. 08 / 984,037 filed on December 2, 1997 and titled "METHOD AND APPARATUS FOR ENCODING MODE SIGNALS FOR USE IN A BINARY SHAPE CODER" (METHOD AND APPARATUS FOR CODING SIGNALS OF MODE FOR USE IN A BINARY ENCODER) . As the signal of binary form, that is to say, the shape information, the texture information can also be encoded in adaptive form by using the interlaced coding and progressive coding techniques to improve its coding efficiency.

As a result of the adaptive coding processes, if the shape information and the texture information are encoded by the progressive coding technique, frame-based motion vectors can be generated according to a motion estimation method based on frame and In another way, field-based motion vectors can be produced based on a field-based motion estimation method. However, the conventional binary coding method uses only the frame-based motion vectors to encode the motion vector of the binary signal. DESCRIPTION OF THE INVENTION Therefore, a primary objective of the invention is to provide a method and an apparatus, to use in adaptively encoding a binary signal by using interlaced and progressive coding techniques, to effectively encode a vector of movement of the signal in binary form when using motion vectors based on field motion and motion vectors. In accordance with one aspect of the present invention, there is provided a method for use in a video signal encoder that encodes a video signal comprising texture information and shape information by selectively employing an interlaced or progressive encoding technique. , to encode a motion vector of the shape information, wherein the shape information is divided into a multiplicity of BABs of M x N binary pixels, and the texture information has a plurality of macro-blocks of the same number of pixels that in the BAB, M and N are positive integers respectively, which comprise the steps of: (a) deciding a type_of_coding (encoding_type) which represents the most effective coding technique between the interleaved and progressive coding techniques for encoding a target BAB; (b) detecting, in response to encoding_type, a motion vector forecaster that corresponds to the target BAB among candidate motion vector predictors that include field-based and box-based motion vectors for texture and shape information; (c) determining the motion vector of the target BAB, based on the detected motion vector predictor; (d) encode the target BAB motion vector. According to another aspect of the present invention, there is provided an apparatus for use in a video signal encoder that encodes a video signal constituted by text information and form information, by selectively using a progressive or interlaced coding technique, to encode a motion vector of the shape information, wherein the shape information is divided into a multiplicity of BABs of M x N binary pixels and the texture information has a plurality of macro-blocks of the same number of pixels as that of the BAB, M and N are positive integers, respectively, comprising: a decision block of encoding_type to determine the most effective coding technique between the interlaced and progressive coding techniques, to encode a target BAB to thereby produce a encoding-type signal; a reconstitution block to provide in response the encondíng__type signal, either the target BAB or background and top field BABs, which are generated by dividing the target BAB according to the interlaced coding technique; an MVP termination block to detect a field-based or frame-based movement vector predictor that corresponds to the target BAB or the background and top field BABs, among the candidate motion vector predictors including field-based motion vectors and based on table for texture and shape information; a movement estimation block to decide the movement vector corresponding to the target BAB when using the detected motion vector forecaster; and a motion vector coding block for encoding the motion vector corresponding to the target BAB. BRIEF DESCRIPTION OF THE DRAWINGS The foregoing and other objects and features of the present invention will be apparent from the following description of preferred embodiments given in conjunction with the accompanying drawings in which: Figure 1 shows a block diagram of a signal coder Of video; Figure 2 is a block diagram of the form coder 200 of Figure 1 according to the present invention; Figure 3 depicts a detailed block diagram of the circuit for motion compensation and motion estimation 260 in Figure 2; Figures 4A to 4C provide a flow chart showing the motion vector coding process according to the present invention; and Figures 5A and 5B illustrate the motion vector forecaster determination process. MODES FOR CARRYING OUT THE INVENTION With reference to Figure 1, a block diagram of a video signal encoder is provided for encoding texture information and shape information for an object in a frame (or VOP). The texture information is expressed by a plurality of macro blocks of M x N pixels, for example, 16 x 16 pixels and the shape information is represented by a multiplicity of binary alpha blocks (BABs) of the same number of pixels as a macro block, M and N are positive integers, respectively. A texture encoder 100 adaptively encodes the texture information coupled to it on a macro block basis by using either a progressive or interlaced coding technique. According to the interlaced coding technique, a macro block is divided into macro blocks of background field and upper field, each macro field block has M / 2 x N pixels, for example 8 x 16 pixels, and the two macro Field blocks are coded on a field-by-field basis. On the other hand, in the technique of progressive coding, a macro block is encoded without any process such as the process of previous division. As a result of the above texture coding processes, the texture encoder 100 provides coded texture information to a data MUX 300 and also supplies texture vector information for texture, which is generated when coding the macro block by using either the technique of progressive or interlaced coding, to a form encoder 200 by an IOL line. Previously, if there is no motion vector corresponding to the macro block, the motion vector information for the texture includes data representing the absence of a motion vector corresponding to the macro block and if there is a motion vector, it includes the vector of movement. More specifically, if the macro block is intercoded using the progressive coding technique, there is a motion vector based on a box corresponding to the macro block. On the other hand, if the macro block is intercoded using the entangled coding technique, there are two field-based motion vectors, ie upper field motion and background field vectors, which correspond to the upper field macro blocks. and of bottom field respectively. Meanwhile, if the macro block is intracoded, its corresponding motion vector is undefined and there is no motion vector. The shape encoder 200 also adaptively encodes the information so that it is fed into a BAB-by-BAB base when using any of the progressive or interleaved coding technique and provides encoded information to the data MUX 300. The MUX of data 300 multiplies the coded texture information that is provided from the texture encoder 100 and the encoded form information provided from the form encoder 200, so as to provide a coded video signal to a transmitter (not shown) for transmission. With reference to Figure 2, a block diagram of the form coder 200 is illustrated in Figure 1 according to one embodiment of the present invention. The shape information is coupled to a circuit for frame mode detection 210 as a current BAB. The frame mode detection circuit 210 examines whether a current BAB coding mode is "all_0" or "all_255". Specifically, the current BAB is divided into T x S pixels, for example 4 4 binary pixels, T and S are positive integers respectively, and therefore the current BAB has 4 sub-blocks, each sub-block contains 4 x 4 pixels primary. If an error between any sub-block of the current BAB and a sub-block of a BAB all_0 is less than or equal to a predetermined threshold, a signal indication Bl of type 1 indicates that the coding mode of the current BAB is " all_0"is provided from the frame mode detection circuit 210 to a frame coding circuit 240, where the BAB all_0 is a BAB where each pixel value is 0. If an error between any sub-block of the BAB current and a sub-block of a BAB all_255 is smaller than or equal to the predetermined threshold, an indication signal Bl of type 2 indicating that the coding mode of the current BAB is "all_255" is provided from the detection circuit of frame mode 210 to the frame coding circuit 240, where the BAB all_255 is a BAB each of whose pixel values is 255. If the coding mode of the current BAB is determined as neither "all_0" nor "all_255", the circuit for mode detection 210 provides the current BAB to an encoding__type 220 decision circuit, a switch 230, and a motion estimation and motion compensation circuit (ME &; MC Motion Estimation and Motion Compensation), respectively 260 through a line L20. The encoding_type decision circuit 220 determines a more effective coding technique between the progressive and interleaved coding techniques for encoding the current BAB that is provided from the frame mode detection circuit 210 by comparing the spatial correlation between rows of pixels in a BAB current-based frame and a BAB current-based field, and produces an encoding_type signal that represents the particular encoding technique. The encoding_type signal is provided to the switch 230, the circuit ME and MC 260 and a multiplexer (MUX) 270. The switch 230 provides the current BAB transmitted by the line L20 to either the frame coding circuit 240 via a line L22 or a field coding circuit 250 on a line L24 in response to the encoding_type signal El. To be more specific, if the encoding_type signal represents the progressive coding technique, the current BAB is coupled to the frame coding circuit 240 and, if otherwise, the current BAB is fed to the field coding circuit 250. Subsequently, either the frame coding circuit 240 or the field coding circuit 250 performs its operation depending on the current BAB is fed. If the current BAB is coupled to the frame coding circuit 240 via the line L22, the frame coding circuit 240 encodes the current BAB in cooperation with the ME and MC 260 circuit using frame-based coding, i.e. progressive coding technique, thereby generating coded data in frame and a frame mode signal corresponding to the current BAB, wherein the frame mode signal represents a coding condition of the data coded in frame and shows one of 7 modes described in Table 1 in the prior art. Specifically, the mode selection between modes 2 and 3, ie the modes "all_0" and "all_255", is determined based only on the indication signal Bl that is provided from the frame mode detection circuit 210. The frame-encoded data containing a coded frame mode signal there is provided to the MUX 270 via a line L42 and the frame mode signal is transferred to the ME and MC 260 circuit via a line L26. In contrast, if the current BAB is fed to the field coding circuit 250 by the line L24, the field coding circuit 250 codes the current BAB in cooperation with the ME and MC 260 circuit when using field-based coding, i.e. the interlaced coding technique, thereby producing the field coded data and a field mode signal corresponding to the current BAB, wherein the field mode signal illustrates a coding condition of the field coded data and represents one of 5 modes described in Table 1 except modes 2 and 3, since modes 2 and 3 are determined only in the frame coding circuit 240 as mentioned above. The field-encoded data containing a coded field-mode signal is supplied to the MUX 270 via a line L44 and the field-mode signal is supplied to the ME and MC 260 circuit by a line L28. The MUX 270 multiplexes the data encoded in frame and field encoded data together with the signal encoding_type El to provide the information in coded form to the data MUX 300 in Figure 1. Previously, the cooperation between any of the frame coding circuit 240 or the field coding circuit 250 and the ME and MC 260 circuit will be illustrated with reference to Figure 3. The ME and MC 260 circuit performs a field-based and frame-based motion estimation procedure in the current BAB based in the current BAB on the L20 line, the encoding_type signal El, the texture vector information in the texture that is provided from the texture encoder 100 in Figure 1 by the IOL line, and any of the frame mode signal on line L26 or the field mode signal on line L28. With reference to Figure 3, a detailed block diagram of the circuit ME and MC 260 in Figure 2 is illustrated in accordance with the embodiment of the present invention. The texture motion vector information that is provided from the texture encoder 100 in Figure 1 is stored in a texture memory MV 269-2. The current BAB coupled to the ME and MC 260 circuit on line L20 is fed to a reconstitution block 262. In response to the encoding_type El signal, the reconstitution block 262 either divides the current BAB into upper field and field BABs background and provides the upper and background field BABs on line L50, or directly transfers the current BAB on line L50, where the upper field BAB preferably contains each row non of the current BAB and the field BAB of background includes each row pair of the current BAB. Either the current BAB or the upper and background field BABs that are sent out from the reconstitution block 262 are provided to a motion estimation block 263 and a reconstruction block 266 along the line L50. Meanwhile, a determination block MVP 261 detects a motion vector forecaster for form (MVP) which corresponds to the current BAB between a plurality of candidate MVPs stored in form memory VM 269-1 and texture memory MV 269- 2 according to predetermined rules in accordance with the present invention in response to the encoding_type signal El. The MVP detection method will be illustrated below with reference to Figures 4A to 4C and Figures 5A and 5B.

According to the present invention, since the shape information and the texture information of the video signal are encoded in an adaptive manner by using the progressive and interlaced coding techniques, the MV memory of form 269-1 also stores information of movement vector in form that is similar to texture vector information in texture stored in texture memory 269-2. That is, if there is no motion vector corresponding to the current BAB, the motion vector information in the form contains data representing the absence of a motion vector corresponding to the current BAB, and if there is a motion vector, it includes the vector of movement as well as movement vector information in texture. With reference to the coding modes of the binary signal shown in Table 1, if the coding mode of the current BAB is determined as one of modes 0, 1, 5 and 6, there is at least one motion vector for the BAB current and if the coding mode becomes one of modes 2, 3 and 4, there is no motion vector for the current BAB since no motion vector is defined for modes 2, 3 and 4. In addition, if the BAB stream is encoded by the progressive encoding technique, the motion vector information in the corresponding form can include a frame-based motion vector and if it is encoded by the interlaced coding technique, it can have two field-based motion vectors , that is, a vector of upper field movement and one of background, corresponding to the upper and background field BABs of the current BAB since the upper and background field BABs are coded on the field-by-field basis. Previously, the motion vector information in the form is determined in a selection block 268 based on the mode signal for each BAB on line L26 or L28 and a motion vector (MV) on a line L55 that can already be generated or not generated by each BAB in the motion estimation block 263. That is, if the coding mode of a BAB is one of modes 2, 3 and 4 and although there is a corresponding motion vector, the vector of movement is not significant. Therefore, in that case, the selection block 268 provides the motion vector information in the form, which means that there is no motion vector to the memory MV of form 269-1. On the other hand, if the BAB coding mode is one of the modes 0, 1, 5 and 6, there is at least one motion vector for the BAB and therefore the motion vector information in the form containing the Motion vector is coupled to memory MV of form 269-1.

According to the previous illustration, frame-based motion vectors and field-based motion vectors stored in shape memory VM 269-1 and texture memory MV 269-2 can be candidate MVPs. Therefore, among the candidate MVPs, the MVP 261 determination block detects either a frame-based MVP or a field-based MVP for the current BAB in response to the encoding_type El signal. If the current BAB is encoded by the technique of progressive coding, ie the signal encoding_type The represents the frame-based encoding, the frame-based MVP for the current BAB, is generated. On the other hand, if the current BAB is encoded by the interlaced coding technique, that is, the encoding_type signal represents the field-based, field-based MVP, including upper field MVPs and background field MVPs corresponding to the BABs. of upper field and background field, respectively, for the current BAB, occurs. Then, the MVP determination block 261 provides the determined MVP, i.e. either the field-based or frame-based MVP to the motion estimation block 263 and a MVD calculation block 264. The motion estimation block 263 determines and a frame-based or field-based motion vector (MV) for the current BAB based on either the frame-based or field-based MVP that is provided from the MVP determination block 261, any of the current BAB or the upper field and background field BABs that are output from the reconstitution block 262 and a previous frame retrieved from a frame memory 267. For example, as described in MPEG-4, The frame-based MV for the current BAB is determined based on its frame-based MVP by the following table-based motion estimation procedure. An MC error is calculated by comparing the BAB indicated by the frame-based MVP and the current BAB. If the calculated MC error is less than or equal to a predetermined threshold for any 4 x 4 sub-blocks, the frame-based MVP is used directly as the frame-based MV and the procedure terminates. However, if the above condition is not satisfied, the candidate motion vectors are searched around the frame-based MVP while 16 x 16 MC errors are calculated by comparing the BABs indicated by the candidate motion vectors and the current BAB. The search range is, for example +/- 16 pixels around the frame-based MVP on both horizontal and vertical directions. Subsequently, a candidate motion vector that minimizes the MC error is taken as the base frame MV of the current BAB. The field-based motion estimation procedure is also performed on each field BAB in the same way as illustrated in the frame-based motion estimation procedure. The MVD of the current BAB, that is, any of the frame-based or field-based MV, is provided to the calculation block MVD 264, the motion compensation block 265 and the selection block 268 for the line L55. The calculation block MVD 264 calculates a motion vector difference for form (MVD) corresponding to the current BAB, the MVD represents a difference between the MV and the MVP, and provides the MVD on a line L30 or L32 depending on the encoding_type of the BAB current. The motion compensation block 265 produces a moving compensated BAB based on the MV derived from the motion estimation block 263 and the previous frame recovered from the frame memory 267 through the use of a conventional motion compensation method; it supplies the BAB compensated by movement on a line and L34 or L36 depending on the encoding__type of the current BAB; and the reconstruction block 266 is provided thereto. The reconstruction block 266 generates a reconstructed BAB by selecting the moving compensated BAB that is provided from the motion compensation block 265 or any of the current BAB or the upper field BABs and in the line L50 derived from the reconstitution block 262 in response to the current BAB mode signal on line L26 or L28. On the other hand, if the mode signal on line L26 that is output from the frame coding circuit 240 represents one of modes 2 and 3, the reconstructed BAB becomes either BAB all_0 or BAB all_255. The reconstructed BAB is supplied on a line either L38 or L40 depending on the current BAB encoding_type and provides the frame memory 267 in order to produce a reconstructed current frame, wherein the reconstructed current frame is stored in the frame memory 267 as the previous box. Briefly, if the current BAB is coded according to the progressive coding technique, it outputs the calculation block MVD 264, the motion compensation block 265 and the reconstruction block 266 are produced in the frame-by-frame basis. box and feed the frame coding circuit 240 for line L30, L34 and L38 respectively and are otherwise generated on the field-by-field basis and fed to the field coding circuit 250 by lines L32, L36 and L40, respectively. Subsequently, the frame coding circuit 240 determines the coding mode of the current BAB and produces the data encoded in frame based on the indication signal Bl, the current BAB that is transferred to it by the switch 230, the MVD based on the frame in line L30, the BAB offset in movement in line L34 and the BAB reconstructed in line L38 in the same way as described in the prior art. Similarly, the field coding circuit 250 decides the coding modes of the upper field and background field BABs corresponding to the current BAB and generates the field coded data corresponding to the upper field BABs and background field BABs. based on the BAB current coupled by switch 230, the field-based MVD on line L32, the motion-compensated BABs that correspond to the upper field and bottom field BABs on line L36 and the reconstructed BAB on the line L40. Next, a method for encoding and determining motion will be illustrated more specifically with reference to Figures 4A to 4C, and Figure 5A and 5B.

In accordance with the present invention, the MVP for the current BAB is chosen from a plurality of candidate MVPs composed of motion vectors to form BABs eg Sl to S3, encircle the current BAB, eg SB, as illustrated in FIG. Figure 5A and motion vectors for macro block texture, for example TI to T3, surrounding the macro block, for example TB corresponding to the current BAB SB as illustrated in Figure 5B. Each of the BABs surrounding the current BAB and the macro blocks surrounding the macro block corresponding to the current BAB already has a frame-based motion vector or a field-based motion vector including upper field and field motion vectors background, depending on your encoding_type. Therefore, in order to determine the frame or field-based MVPs of the current BAB, the candidate MVPs are traversed according to the priority determined by MPEG-4 and the encoding_type of the current BAB. According to MPEG-4, the BABs Sl to S3 surrounding the current BAB SB have a priority in the order of Sl, S2 and S3 as illustrated in Figure 5A and similarly, the macro blocks TI to T3 that surround the macro TB block are traversed in the order of IT, T2 and T3. The dependent priority of the encoding_type will be described when referring to Figures 4A to 4C. In Figure 4A, in the SIO stage, it is verified whether the current BAB encoding_type determined in the encoding_type decision circuit 220 in Figure 2 is the type of progressive coding. If the encoding_type of the current BAB is determined as the progressive encoding type in the SIO stage, among the candidate MVPs, the frame-based motion vectors are traversed before the field-based motion vectors. In order to determine the MVP based on the current BAB table, the procedure first proceeds to step S20. In step S20, it is detected if there exists a motion vector based on frame appropriate for the form (MVS), between the candidate MVPs in the order of Sl to S3 in Figure 5A. If there are table-based MVSs, the procedure passes from step S70 and the detected frame-based MV_s are determined as the frame-based MVP of the current BAB in step S70. The procedure continues to step S90 thereafter. If otherwise in step S30, it is observed if there is an MV = field-based appropriate among the candidate MVPs in the order of Sl to S3 in Figure 5A. As a result of the detection process in step S30, if there are field-based MVSs including upper field movement and background field vectors, the procedure proceeds to step S80. In step S80, an average of the upper field movement and background field vectors included in the field-based MVS is calculated and the average is determined as the frame-based MVP of the current BAB.

The procedure proceeds to step S90 subsequently. If no field-based MVS are detected in step S30, in step S40, it is checked whether there is a motion vector based on appropriate frame for texture (MBT) between the candidate MBPs in the order of IT to T3 in the Figure 5B. If there is MVT, based on frame, frame-based MVT is determined as the frame-based MVT of the current BAB in step 70 and, otherwise, the procedure advances to step S50. With reference to step S50, it is verified if there is an appropriate field-based MVT between the candidate MVPs in the same sequence as in step S40. If there is the field-based MVT that includes upper field movement and background field vectors, the procedure proceeds to step S80. In step S80, an average of the upper field movement and background field vectors included in the field-based MVT is calculated and the average is determined as the frame-based MVP of the current BAB. The procedure proceeds to step S90 subsequently. On the other hand, as a result of the detection processes in steps S20 to S50, if it is revealed that there is no frame-based or field-based motion vector among the candidate MVPs, the frame-based MVP of the current BAB is set to zero in step S60. Steps S20 to S80 which determine the frame-based MVP of the current BAB are performed in the MVP determination block 261 in Figure 3. Then, the procedure goes to step S90 in Figure 4C and in step S90, the MV Based on the current BAB table, it is decided based on the MVP based on the current BAB table according to MPEG-4 above, as was done with the motion estimation block 263 shown in Figure 3. In step S100, the MVD based on the current BAB table is calculated by subtracting the MVP based on the MV table based on the current BAB table. Then, the frame-based MVD is transferred to the frame coding circuit 240 via the line L30 as illustrated in Figure 2 in step S110 and this coding and motion vector determination procedure ends.

On the other hand, the SIO stage in Figure 4, if the encoding__type of BAB stream is not the progressive encoding type, ie if the current BAB is encoded by the interlaced coding technique, the field-based motion vectors are traversed before frame-based motion vectors and the procedure goes to step S120 in Figure 4B. In step S120, it is detected if there is an MVS based on appropriate field including upper field movement and background field vectors between the candidate MVPs in the order of Sl to S3 in Figure 5A. If there are field-based MVSs, the procedure proceeds to step S160. In step S160, the field-based MVP that includes the top-field and bottom-field MVPs is determined based on the field-based MVPs, where the top-field and bottom-field MVPs correspond to the BABs of upper field and background field of the current BAB, respectively. In accordance with the embodiment of the present invention, the upper field movement vector of the field-based MVS is determined as the upper field MVP and the background field movement vector is decided as the background field MVP. of the current BAB. According to another embodiment of the present invention, the field-based field movement vector of the field-based MVS is determined as the upper field and background field MVPs of the current BAB. On the other hand, if the appropriate field-based MVb is not chosen in step S120, in step S130 it is examined whether an MVS based on an appropriate frame exists between the candidate MVPs in the order of Sl to S3. If there is frame-based MVS, the frame-based MVS is decided as the upper field and background field MVPs of the current BAB at S160 and otherwise, the procedure proceeds to step S140. In step S140, it is checked whether there is a field-based MVT including upper field movement and background field vectors between the candidate MVPs in the order of TI to T3 in Figure 5B. If it is determined that it is the field-based MVT, the Field-based MVP of the current BAB is determined in step S160 in the same way as in determining the field-based MVP when using the field-based MVS as shown above. On the other hand, if the field-based MVT is not detected in step S140, in step S150 it is checked if there is a frame-based MVT among the candidate MVPs in the same sequence as in step S140. If the MVT is based on a frame, the frame-based MVT is determined as the upper field and background field MVPs of the current BAB in step S160. Then, the procedure proceeds to step S180 in Figure 4C. As a result of steps S120 to S150, if it is discovered that there is no field-based or frame-based motion vector among the candidate MVPs, the field-based MVP of the current BAB is set to zero in step S170. Then, the procedure proceeds to step S180 in Figure 4C. In step S180, the field-based MV of the current BAB is decided on the field-by-field basis according to the MPEG-4 above, in order to produce the upper field and background field VMs of the BAB current. If the field-based MV is determined at the stage S180, the field-based MVD is calculated by using the field-based MV and MVP of the current BAB on the field-by-field basis in step S190. In step S200, the field-based MVD is transferred to the field coding circuit 250 via line L32 as illustrated in Figure 2 and this coding and motion vector determination procedure is terminated. While the present invention has been described with respect to certain preferred embodiments only, other modifications and variations may be made without departing from the spirit and scope of the present invention as set forth in the following claims.

Claims (16)

1. A method for using in a video signal encoder that encodes a video signal comprising texture information and shape information by selectively using a progressive or interlaced coding technique, to encode a motion vector of shape information, wherein the shape information is divided into a multiplicity of BABs of M x N binary pixels and the texture information has a plurality of macro blocks of the same number of pixels as that of BAB, M and N are positive integers, respectively comprising the steps of: (a) deciding an encoding__type representing a more effective coding technique between progressive and interlaced coding techniques to encode a target BAB; (b) detecting, in response to the encoding_type, a motion vector forecaster that corresponds to the target BAB among candidate motion vector predictors including frame-based and field-based motion vectors for shape and texture information; (c) determining the target BAB motion vector based on the detected motion vector predictor; and (d) encoding the target BAB movement vector. The method according to claim 1, characterized in that the candidate motion vector predictors include frame-based and field-based motion vectors for BABs surrounding the target BAB and have been field-based and field-based motion vectors. table and previously encoded for macro blocks surrounding the macro block corresponding to the target BAB and have also been previously encoded, where each field-based motion vector has upper field movement and background field vectors that correspond to field blocks upper and background field respectively of a BAB or a macro block. 3. The method according to claim 2, characterized in that if the type of progressive coding is decided as an objective BAB encoding_type in step (a), step (b) detects the forecaster of movement vector that travels the predictors of motion vector candidates in the order of frame-based movement vectors for shape information, field-based motion vectors for shape information, frame-based motion vectors for texture information and frame vectors Field-based movement for texture information. The method according to claim 3, characterized in that in step (b), if one of the field-based motion vectors is chosen to detect the motion vector forecaster, step (b) includes the steps of : (bll) calculate an average of the upper field movement and background field vectors of the selected field-based motion vector; and (bl2) determine the average as the forecaster of motion vector. The method according to claim 2, characterized in that if the type of interleaving coding is decided as the encoding_type of the target BAB in step (a), step (b) includes the steps of: (b21) dividing the BAB objective in upper field BABs and background field; and (b22) determining the motion vector forecaster that has upper field movement vector and background field forecasters that correspond to the upper field and bottom field BABs respectively, which traverse the candidate motion vectors in the order of field-based motion vectors for shape information, frame-based motion vectors for shape information, field-based motion vectors for texture information and frame-based motion vectors for information of texture. 6. The method according to claim 5, characterized in that in step (b22), if one of the frame-based motion vectors is chosen to detect upper field movement vectors and background field predictors, step (b22) assigns the motion vector based on select box to the upper field BABs and of background field as foreground vectors of top field movement and background field, respectively. The method according to claim 5, characterized in that in step (b22), if one of the field-based motion vectors is chosen to detect upper field movement vector and background field predictors, the stage (b22) determines the upper field movement vector of the selected field-based motion vector as the upper field motion vector forecaster and decides the background field motion vector as the background field motion vector forecaster . The method according to claim 5, characterized in that in step (b22), if one of the field-based motion vectors is chosen to detect the forecaster of upper field movement and background field, the stage (b22) allocates the background field movement vector of the selected field-based movement vector to the upper field and background field BABs of the target BAB and the upper field and background field movement vector predictors, respectively . 9. An apparatus for use in a video signal encoder that encodes a video signal comprising texture information and form information, by selectively using a progressive or interlaced coding technique, to encode a vector of information movement of form, wherein the shape information is divided into a multiplicity of BABs of M x N binary pixels and the texture information has a plurality of macro blocks of the same number of pixels as those of a BAB, M and N are positive integers, respectively, comprising: means for determining a more effective coding technique between progressive and interleaved coding techniques for encoding a target BAB to thereby produce an encoding_type signal; means for providing, in response to the encoding_type signal, any target BAB or a higher field and background field BABs, which are generated by dividing the target BAB according to the interlaced coding technique; MVP determination means to detect a field-based or frame-based motion vector predictor that corresponds to either the target BAB or the upper field and background field BABs among candidate motion vector predictors including motion vectors based in frame and field-based for texture and shape information; means for deciding the motion vector that corresponds to the target BAB by using the detected motion vector forecaster; and means for encoding the motion vector corresponding to the target BAB. The apparatus according to claim 9, characterized in that the candidate motion vector predictors include frame-based and field-based motion vectors that correspond to BABs that encircle the target BAB and that have been previously encoded and motion vectors frame-based and field-based that correspond to macro blocks surrounding the macro block that corresponds to the target BAB and have also been previously encoded, where each field-based motion vector has upper field movement and background field vectors which correspond to upper field blocks and background field respectively of a BAB or a macro block. The apparatus according to claim 10, characterized in that the means for MVP determination choose the box-based movement vector forecaster when traversing the candidate motion vector predictors in the order of frame-based motion vectors. shape information, field-based motion vectors for shape information, frame-based motion vectors for texture information and field-based motion vectors for texture information. The apparatus according to claim 11, characterized in that if the table-based motion vector forecaster is detected among the field-based motion vectors for the texture and shape information, the MVP determination means calculate an average of the upper field movement vectors and the background field vector of movement based on detected field and determine the average as the prognostic vector of motion based on frame. The apparatus according to claim 10, characterized in that the means for MVP determination choose the field-based motion vector forecaster including upper field movement vector and background field forecasters that correspond to the upper field BABs and background fieldrespectively when traversing candidate motion vectors in the order of field-based motion vectors for shape information, frame-based motion vectors for shape information, field-based motion vectors for texture information and frame-based motion vectors for texture information. 14. The apparatus according to claim 13, characterized in that if the field-based motion vector predictor is detected between the frame-based motion vectors for the texture and shape information, the MVP determination means allocates the vector of motion. frame-based movement detected to the upper field BABs and background field BABs as the upper field movement vector and background field forecasters respectively. 15. The apparatus according to claim 13, characterized in that if the field-based motion vector predictor is detected between the field-based motion vectors for the texture and shape information, the MVP determination means decide the vector of upper field movement of motion vector based on field detected as the forecaster of upper field movement vector and its background field motion vector as the forecaster of background field movement vector. 16. The apparatus according to claim 13, characterized in that if the field-based motion vector predictor is detected between the field-based motion vectors for the texture and shape information, the MVP determination means allocates the Background field motion of the field-based movement vector detected to the background BABs of the target BAB as foreground and top-field motion vector predictors, respectively. SUMMARY OF THE INVENTION An apparatus encodes a video signal composed of texture information and form information, by selectively using a progressive encoding technique and an interlaced one. The shape information is divided into a multiplicity of BABs of M x N pixels having one of a first and a second binary value and the texture information has a plurality of macro blocks of the same number of pixels as BAB. Specifically, the apparatus determines a motion vector of a target BAB of shape information, based on a corresponding motion vector predictor selected among candidate motion vectors including frame-based and field-based motion vectors for surrounding BABs The target BAB and previously have been coded and frame-based and field-based motion vectors for macro blocks surrounding the macro block corresponding to the target BAB and have also been previously coded according to given priority, and encode the movement vector of the BAB objective. 18655 1/7