Abstract: A machine accesses view maps that each corresponds to a different point on a surface of a three-dimensional object. The machine generates vectors of coefficients by calculating a corresponding transform of each view map. The vectors specify a corresponding coefficient for each basis function among a group of basis functions. For each basis function in the group, the machine generates a spatial representation of coefficients specified for that basis function across all of the accessed view maps, and then quantizes the spatial representation of coefficients for that basis function. The quantized spatial representation forms part of a group of quantized spatial representations of coefficients that corresponds to the accessed view maps. The machine then provides the group of quantized spatial representations of coefficients as a compressed version of the view maps. A decoder device can then approximate the view maps based on the group of quantized spatial representations.

Abstract: A method for fast multiple reference frame motion estimation, which is used to perform motion estimation between a current frame and reference frames Fn?1, Fn?2, . . . , Fn?k. The method performs a special block matching to find a plurality of motion vectors for blocks of each frame with respective to a previous frame and then composes motion vector(s) of the current frame referring to the reference frame Fn?(k?1) and a motion vector of the reference frame Fn?(k?1) referring to the reference frame Fn?k into composed motion vectors of the current frame referring to the reference frame Fn?k for the block in the current frame. The method selects a composed motion vector with a minimum of cost function from the composed motion vectors produced when composing the motion vectors and then finely adjusts the composed motion vector selected, thereby obtaining an adjusted motion vector.

Abstract: Methods and systems for fast variable block-size motion estimation based on merging and splitting procedures for AVC video encoding are disclosed. The methods take advantage of the correlation of the Motion Vectors (MVs) of the different block-size modes to achieve a good computation reduction. Considering that the smaller the block-size difference between two block-sizes is, the more correlations between their MVs can be expected, the methods use a 8*8 block as an initial processing block to calculate prediction MVs. The prediction MVs are then used in predicting MVs for other block-sizes.

Abstract: A method for fast macroblock mode decision is disclosed. The method includes: (A) determining if a motion cost at the origin (0, 0) or a prediction motion vector (PMV) for a 4n*4n macroblock is smaller than a first threshold; (B) if the motion cost is smaller than the first threshold, determining a macroblock mode as 4n*4n and ending the method; (C) if the motion cost is not smaller than the first threshold, using an adaptive diversity search strategy to perform motion estimation on four 2n*2n blocks associated with the 4n*4n macroblock; (D) determining if all motion costs of the four 2n*2n blocks in step (C) are smaller than a second threshold; and (E) if step (D) determines that the motion costs of the four 2n*2n blocks are smaller than the second threshold, determining the macroblock mode as 2n*2n and ending the method.

Abstract: Methods and systems for fast variable block-size motion estimation based on merging and splitting procedures for AVC video encoding are disclosed. The methods take advantage of the correlation of the Motion Vectors (MVs) of the different block-size modes to achieve a good computation reduction. Considering that the smaller the block-size difference between two block-sizes is, the more correlations between their MVs can be expected, the methods use a 8*8 block as an initial processing block to calculate prediction MVs. The prediction MVs are then used in predicting MVs for other block-sizes.

Abstract: A method for fast macroblock mode decision is disclosed. The method includes: (A) determining if a motion cost at the origin (0, 0) or a prediction motion vector (PMV) for a 4n*4n macroblock is smaller than a first threshold; (B) if the motion cost is smaller than the first threshold, determining a macroblock mode as 4n*4n and ending the method; (C) if the motion cost is not smaller than the first threshold, using an adaptive diversity search strategy to perform motion estimation on four 2n*2n blocks associated with the 4n*4n macroblock; (D) determining if all motion costs of the four 2n*2n blocks in step (C) are smaller than a second threshold; and (E) if step (D) determines that the motion costs of the four 2n*2n blocks are smaller than the second threshold, determining the macroblock mode as 2n*2n and ending the method.

Abstract: A method for fast multiple reference frame motion estimation, which is used to perform motion estimation between a current frame and reference frames Fn?1, Fn?2, . . . , Fn?k. The method performs a special block matching to find a plurality of motion vectors for blocks of each frame with respective to a previous frame and then composes motion vector(s) of the current frame referring to the reference frame Fn?(k?1) and a motion vector of the reference frame Fn?(k?1) referring to the reference frame Fn?k into composed motion vectors of the current frame referring to the reference frame Fn?k for the block in the current frame. The method selects a composed motion vector with a minimum of cost function from the composed motion vectors produced when composing the motion vectors and then finely adjusts the composed motion vector selected, thereby obtaining an adjusted motion vector.

Abstract: Methods and systems for fast variable block-size motion estimation based on merging and splitting procedures for AVC video encoding are disclosed. The methods take advantage of the correlation of the Motion Vectors (MVs) of the different block-size modes to achieve a good computation reduction. Considering that the smaller the block-size difference between two block-sizes is, the more correlations between their MVs can be expected, the methods use a 8*8 block as an initial processing block to calculate prediction MVs. The prediction MVs are then used in predicting MVs for other block-sizes.

Abstract: Methods and systems for providing a video stream from a server to a client over a network include a memory for storing a forward-encoded bit-stream and a reverse-encoded bit-stream for a video data. The forward-encoded bit-stream includes I-frames encoded without inter-frame dependencies and P-frames encoded depending on forward-direction preceding frames, and the reverse-encoded bit-stream includes I-frames and P-frames encoded depending on reverse-direction preceding frames. When the server receives a request with a video cassette recording (VCR) function from the client, the server reads out and transmits frames selectively from among the first, second, third, and fourth frames in accordance with the request. The server can select the closest I-frame to a requested frame in either bit-stream for a fast-mode play or a random-access play, and switch the bit-streams to use subsequent P-frames in a different direction than that of the closest I-frame.

Abstract: A transcoder for transcoding digital video signals includes a decoder and an encoder. In the decoder, an end-of-block (EOB) position of an incoming block received by the decoder is determined and a discrete cosine transform (DCT) block type is determined based on the determined EOB position. A reduced number of DCT coefficients is computed in a subsequent inverse DCT computation based on the DCT block type. In the encoder, if the incoming block is intercoded, no DCT coefficients are computed after the EOB of the incoming blocks is performing a DCT. Further, in the encoder when the incoming block is intercoded, an algorithm is applied to predict which DCT coefficients may become zero after a subsequent quantization operation, and only DCT coefficients that may not become zero are computed in performing the DCT.

Abstract: Methods and systems for providing a video stream from a server to a client over a network include a memory for storing a forward-encoded bit-stream and a reverse-encoded bit-stream for a video data, The forward-encoded bit-stream includes I-frames encoded without inter-frame dependencies and P-frames encoded depending on forward-direction preceding frames, and the reverse-encoded bit-stream includes I-frames and P-frames encoded depending on reverse-direction preceding frames. When the server receives a request with a video cassette recoding (VCR) function from the client, the server reads out and transmits frames selectively from among the first, second, third, and fourth frames in accordance with the request. The server can select the closest I-frame to a requested frame in an either bit-stream for a fast-mode play or a random-access play, and switch the bit-streams to use subsequent P-frames in a different direction than that of the closest I-frame.

Abstract: A transcoder for transcoding digital video signals includes a decoder and an encoder. In the decoder, an end-of-block (EOB) position of an incoming block received by the decoder is determined and a discrete cosine transform (DCT) block type is determined based on the determined EOB position. A reduced number of DCT coefficients is computed in a subsequent inverse DCT computation based on the DCT block type. In the encoder, if the incoming block is interceded, no DCT coefficients are computed after the EOB of the incoming blocks is performing a DCT. Further, in the encoder when the incoming block is intercoded, an algorithm is applied to predict which DCT coefficients may become zero after a subsequent quantization operation, and only DCT coefficients that may not become zero are computed in performing the DCT.

Abstract: Methods and systems for generating motion vectors for re-encoding video signals are disclosed. The motion vector is determined by the sum of a base motion vector and a delta motion vector. In the case of no frame-skipping, the base motion vector is the incoming motion vector. In the case of frame skipping, the base motion vector is the sum of the motion vectors of the incoming signal since the last re-encoded frame and the current frame. The delta motion vector is optimized by a minimum Sum of the Absolute Difference by searching over a smaller area than if searching for a new motion vector without a delta motion vector. These methods and systems may be used to improve re-encoding digital video signals.

Abstract: Methods and systems for generating motion vectors for re-encoding video signals are disclosed. The motion vector is determined by the sum of a base motion vector and a delta motion vector. In the case of no frame-skipping, the base motion vector is the incoming motion vector. In the case of frame skipping, the base motion vector is the sum of the motion vectors of the incoming signal since the last re-encoded frame and the current frame. The delta motion vector is optimized by a minimum Sum of the Absolute Difference by searching over a smaller area than if searching for a new motion vector without a delta motion vector. These methods and systems may be used to improve re-encoding digital video signals.

Abstract: An apparatus and method for combining video signals from multiple users in a continuous presence video conferencing system. Video signals are received from each of a number of different system users. The signals include input frames each having a number of different groups-of-blocks (GOBs). The GOBs correspond to hierarchically organized video data and header information. The input frame GOBs are used to form a composite output video signal suitable for transmission to each of the system users. An output frame of the composite output signal incorporates at least one of the GOBs from the input frames of each of the users and less than all of the GOBs from the input frame of at least one of the users. The resulting output video signal provides flexible continuous presence video conferencing in which video signals from multiple users may be simultaneously displayed to all users in a selected divided-screen format.

Abstract: Four QCIF video input signals generated by participants in a multiparty video conference are combined in the coded domain to produce a merged CIF video output signal. When the CIF video output signal is transmitted back to each party's video terminal, a combined 2.times.2 image is displayed. A video signal combiner (700) combines the input video signals in the coded domain by time-division multiplexing the inputs at the GOB level. In order to maintain frame synchronization between the inputs which may be arriving at different frame repetition rates, the combiner stores the inputs in buffers (706-709) and processes (710) the temporal reference (TR) numbers associated with each frame in each input before merging the GOBs from each input in accordance with the processed TR number. Specifically, the TR of each input is offset by an initial TR number associated with each input and determined at turn-on. The offset TR in each input is then mapped onto a scale of TR numbers that is common to each input.

Abstract: A variable-length codeword encoder is disclosed which produces 8-bit output segments for storage in a buffer (23) for subsequent transmission over a transmission channel (24). The encoder includes two memory tables (15, 16), which produce in response to each input symbol to be encoded, a variable-length codeword and an a codeword length. An accumulator (31, 33) accumulates, modulo-8, the successive codeword lengths, producing a carry signal during any clock cycle in which eight or more bits codeword bits are accumulated. At each clock cycle, the variable-length codeword is input to the parallel inputs of a cross bar shift control circuit (30). This shift control circuit produces a 16-bit output in which the input word is embedded.

Abstract: Method and system for high order conditional entropy coding utilizing an incremental-tree-extension technique to design a conditional tree for the coding. For complexity reduction, code table reduction and non-uniform quantization of conditioning symbols or pixels is provided. A pattern matching technique is provided for fast conditioning state extraction, and a multistage pipelined structure is provided to handle the case of a large number of conditioning pixels. Using the complexity reduction techniques and the hardware structures, it is possible to implement practical high order conditional entropy codecs using current low-cost VLSI (Very Large Scale Integration) technology.

Abstract: A variable-length decoder is disclosed which includes a decoder part (101) and an interface part (102). The decoder part includes two decoder latches (110 and 111) which store consecutive sequences of bits, which are each equal in length to the maximum-length codeword. A decoder barrel shifter (109) provides an output decoding window of a subsequence of bits in the two latches equal in length to the maximum-length codeword. This subsequence is supplied to a memory device (116) which provides a corresponding decoded codeword output and codeword-length output for each subsequence that begins with the first bit of a variable-length to be decoded. At each clock cycle, the same subsequence is supplied to the input of the first decoder latch and the decoder barrel shifter is shifted by the codeword-length of the previous word so that the decoder barrel shifter output subsequence always begins with the first bit of the word to be decoded. The interface part supplies the second decoder latch.

Abstract: A variable-length decoder is disclosed in which a received variable-word-length encoded bit stream is input to a buffer (102) and read out in parallel sequences equal in length to the maximum length codeword. These sequences are read into cascaded latches (105, 107). The cascaded sequences in both latches are input to a barrel shifter (109) which provides from its multi-bit input, a sliding decoding window to a table-lookup memory device (112). A control signal directly shifts the position of the decoding window of the barrel shifter as each codeword is detected. To detect each codeword, the initial bits in the decoding window are compared with the codeword entries in the table-lookup memory. When a codeword is detected, the corresponding stored codeword length is accumulated (120) with previously accumulated codeword lengths to produce the control signal which directly shifts the decoding window by the number of bits in the just decoded word.