Daewoo Electronics Co., Ltd.

Seok-Won Han

Jin-Hun Kim

A video signal encoder codes a video signal including texture information and shape information on each of macroblocks, a macroblock having 16&times;16 pixels and being dividable into 4 number of equal-sized DCT-blocks. The encoder produces encoded shape information by encoding the shape information on a processing macroblock and generates a reconstructed shape information by decoding the encoded shape information. Then, a DCT_type of the processing macroblock is determined based on the reconstructed shape information and the texture information on the processing macroblock to thereby provide DCT_type information representing the DCT_type. Once the DCT_type of the processing macroblock is decided, the encoder generates encoded texture information by adaptively encoding the texture information through the use of a progressive or an interlaced coding technique in response to the DCT_type information. Furthermore, the encoder generates CBPY information by using the encoding results of the texture information and provides a data stream to be transmitted in sequence of the encoded shape information, the encoded CBPY information, the DCT_type information and the encoded texture information.

Method and apparatus for encoding a video signal

A method restores a binary shape signal which has been downsized through the use of a sub-sampling or a down-sampling technique. The downsized binary shape signal contains a plurality of reference lines, wherein each reference line includes one or more segments having successive object pixels therein, and the binary shape signal is restored based on the downsized binary shape signal. Once interpolation lines corresponding to the reference lines are produced based on the number of segments on each of the reference lines, positions of the segments, and the number of object pixels on each of the segments, and then the restored binary shape signal is provided by alternately combining each of the interpolation lines and each of the reference lines.

Method for producing a restored binary shape signal based on an

In a method for encoding a binary alpha block (BAB) within a current frame based on the current frame and a previous frame, even-numbered horizontal lines of the BAB are sampled to generate a first block and even-numbered vertical lines of the first block are sampled to generate a first sample block. And then, a reconstructed first block and a reconstructed BAB are generated by inserting reconstructed vertical and horizontal lines sequentially at locations of odd-numbered vertical and horizontal lines thereof, respectively. In the above, binary values of first and second binary pixels of the reconstructed vertical and horizontal lines are made by using corresponding first and second context values based on predetermined probability tables, respectively, wherein the corresponding first and second context values are calculated by using the context values of binary pixels within the previous frame located at identical positions to the first and second binary pixels, respectively; and context values of binary pixels located at surrounding positions of the first and the second binary pixels except a binary pixel below the first binary pixel and a binary pixel right to the second binary pixel, respectively.

Method for encoding a binary shape signal

A binary shape encoding apparatus encodes a binary alpha block (BAB) consisting of M.times.N binary pixels, M and N each being a positive even integer. The apparatus samples every other horizontal lines of the BAB to generate a first block starting from either a first or a second horizontal line of the BAB and then samples every other vertical lines of the first block to generate a first sample block as a base layer also starting from either a first or a second vertical line of the first block, wherein the first horizontal and vertical lines are the topmost horizontal and the leftmost vertical lines in the BAB and the first block, respectively. Then, the apparatus encodes the first sample block to thereby produce coded base layer data and performs enhancement layer encoding based on the BAB, the first block and the first sample block to thereby provide coded horizontal and vertical enhancement layer data.

Method and apparatus for binary shape encoding

A method for encoding a video signal including a video object plane(VOP) and object information thereof, the VOP having an object therein, divides the VOP into a plurality of blocks and detecting boundary blocks and object blocks, wherein each boundary block includes therein background pixels and object pixels and each object block includes therein object pixels only, the background and the object pixels representing pixels residing outside and inside the object, respectively. Then the object information is converted into a set of object symbols and each object symbol is added to a background pixel included in a boundary block to thereby produce processing blocks, the processing blocks including the object blocks and one or more processed boundary blocks having therein the object symbols and the remaining unprocessed boundary blocks. The processing blocks are encoded to thereby generate an encoded video signal.

Method and apparatus for encoding object information of a video object

A method for scalably inter-contour coding a video signal including a previous and a current frames, wherein each of the previous and the current frames contains a contor, first generates a predicted contour by motion estimating and compensating a previous base layer, widens the predicted contour by a predetermined threshold D.sub.max (j), generates (j)th primary and secondary vertices, and reconstructs a current base layer, j being 0. Thereafter, the method widens the predicted contour by the predetermined threshold D.sub.max (j+1), generates (j+1)st primary and secondary vertices, and updates the (j)th primary and secondary vertices. Then, the (j)th primary and secondary vertices are encoded, and encoded primary and secondary vertex information is formatted in a predetermined way. Finally, the value of j is increased by 1, and the process from widening the predicted contour by D.sub.max (j+1) to formatting the encoded primary and secondary vertex infromation are repeated until j becomes N, N being a positive integer.

Scalable inter-contour coding method and apparatus

A method for encoding mode signals of a target block of M.times.N pixels having a binary value `0` or `255` is provided. If the target block is determined as neither 'all.sub.-- 0' nor 'all.sub.-- 255', either a frame-based coding or a field-based coding is selected to generate a coding mode signal, wherein the frame-based coding represents the target block is encoded on an M.times.N pixels basis and the field-based coding represents the target block is encoded on an M/2.times.N pixels basis, and a base mode of the target block based on the 'all.sub.-- 0', 'all.sub.-- 255' and/or the coding mode signal. If the frame-based coding is selected, a frame mode is generated. On the contrary, if the field-based coding is selected, the frame is divided into a top and a bottom fields so that a top mode for the top field and a modified bottom mode for the bottom field are obtained, wherein a bottom mode is modified by the top mode to be the modified bottom; and the top mode and the modified bottom mode are attached to be a field mode.

Method and apparatus for encoding mode signals for use in a binary shape

In a mode coding method in encoding a binary alpha block (BAB) within a current frame based on the current frame and a previous frame including a plurality of BAB's, the BAB within the current frame is sub-sampled to generate a sample block and then a reconstructed BAB is reconstructed based on the sample block. In an intra mode, after evaluating the difference between the reconstructed BAB and the BAB, if there is no difference therebetween, a first mode signal is generated and if otherwise, a second mode signal is generated and at the same time a second error data block representing the difference is generated. In an inter mode, besides performing the process in the intra mode described above, a BAB most similar to the BAB within the current frame among the BAB's within the previous frame is detected as a predicted BAB and then the difference between the BAB and the predicted BAB is evaluated. And if there is no difference therebetween, a third mode signal is generated and if otherwise, a fourth mode signal is generated and at the same time a fourth error data block representing the difference is generated. If two or more mode signals among the first to fourth signals are generated, one mode signal among them is selected as a selected mode signal according to a predetermined selection rule.

Mode coding method in a binary shape encoding

A video signal encoder codes texture information of a video signal which includes the texture information and shape information on each of macroblocks. The encoder generates a predicted coded block pattern type (CBPY) bit-number based on the shape information on a target macroblock. In the meantime, a DCT.sub.-- type of the target macroblock is determined based on the texture information on the target macroblock, wherein the DCT.sub.-- type represents a more effective coding technique between a progressive and an interlaced coding techniques for encoding the texture information, thereby producing DCT.sub.-- type information representing the DCT.sub.-- type. In response to the DCT.sub.-- type, the texture information on the target macroblock is transformed and CBPY information for the target macroblock is generated based on the transformed texture information. Then, the CBPY information is encoded by using the predicted CBPY bit-number and the transformed texture information is encoded to thereby produce encoded texture information.

A contour encoding apparatus encoding a contour of an object in a current frame based on a previous frame first generates a global motion vector between the current and the previous frames and calculates the number of contours in the previous frame. Then, the apparatus determines, for each of current contours, an optimum contour among the previous contours and detects a local motion vector between the optimum and the current contours. Once the global and the local motion vectors are determined, a predicted contour corresponding to each of the current contours is generated by shifting the optimum contour by the global and the local motion vectors. By using the predicted and the current contours, an error representing a difference therebetween is extracted. Subsequently, flag signals describing how close each of the previous contours is to a corresponding current contour are produced based on the number of the previous contours and contour encoding information corresponding to each of the current contours and sequences of the errors and the local motion vectors for the current frame are outputted based on the contour encoding information corresponding to each of the current contours. Finally, the flag signals and the sequences of the errors and the local motion vectors are multiplexed and then the multiplexed data is transmitted as an encoded contour data.

Method and apparatus for encoding a contour of an object based on a

A method for encoding mode signals of a target block of M.times.N pixels having a binary value `0` or `255` is provided. If the target block is determined as neither 'all.sub.-- 0' nor 'all.sub.-- 255', the target block is divided into a top field and a bottom field, wherein the top field contains every odd row of the target block to have M/2.times.N pixels and the bottom field contains every even row of the target block to have M/2.times.N pixels. After encoding the top field, the method encodes the bottom field based on the top field to generate modified bottom field-coded data and a modified bottom mode.

Growing community of inventors

Seoul, South Korea

Seok-Won Han