Abstract: The present invention relates to a method that uses a motion vector of a predictive video frame of a sub-layer to encode a video signal and decode encoded video data. The method encodes a video signal using a preset method to a bit stream of a base layer while encoding the video signal using a scalable MCTF method to a bit stream of an enhanced layer. When an arbitrary frame of the video signal is encoded, information, enabling at least one vector, derived from a first motion vector of a first block included in the bit stream of the base layer in the same direction as the first motion vector, to be used as a motion vector of an image block in the arbitrary frame, is recorded in the bit stream of the enhanced layer. The first motion vector is directed in the same direction as a temporal direction from the arbitrary frame to the first block.

Abstract: Error concealment is used to hide the effects of errors detected within digital video information. A novel spatial error concealment technique is disclosed for use when the error concealment mode decision determines that spatial error concealment should be used for reconstruction. The novel spatial error concealment technique divides a corrupt macroblock into multiple regions, such as, a corner region, a row adjacent to the corner region, a column adjacent to the corner region, and a remainder main region. Those regions are then reconstructed and information from earlier reconstructed regions may be used in later reconstructed regions. Finally, a macroblock refreshment technique is disclosed for preventing error propagation from harming non-corrupt inter-blocks. Specifically, an inter-macroblock may be ‘refreshed’ using spatial error concealment if there has been significant error caused damage that may cause the inter-block to propagate the errors.

Abstract: A method and system of performing real-time video superresolution. A decoder receives a data stream representing a low resolution video and including global motion vectors relating to image motion between frames of the low resolution video. The decoder uses the global motion vectors from the received data stream and multiframe processing algorithms to derive a high resolution video from the low resolution video. The sharpness of the high resolution video may be enhanced.

Abstract: There is provided a video encoding apparatus allowing for enhanced video encoding speed according to the H.264 video coding standard. The video encoding apparatus allows the memories included in the video encoding apparatus to be shared by a plurality of elements through the rearrangement and the structural change of the memories considering an efficient hierarchical motion estimation algorithm. Therefore, the video encoding apparatus has the effects of reducing the amount of transmitted and received data between the frame memory and the video encoding apparatus and enhancing video encoding speed.

Abstract: A video encoding method and apparatus, and a video decoding method and apparatus, which are capable of improving efficiency of encoding a luminance component video signal by predicting information on various encoding modes of the luminance component video signal by using a result obtained by encoding a chrominance component video signal, are provided. Accordingly, a block mode and an intra-prediction mode of the luminance component video signal are predicted from a previously encoded and recovered chrominance component video signal, and the predicted block mode and the predicted intra-prediction mode are used to encode the luminance component video signal.

Abstract: A method for video encoding is disclosed. The method generally includes the steps of (A) generating first sub-pel data for at least one of (i) a motion estimation and (ii) a mode decision by first filtering reference data and (B) generating second sub-pel data for a motion compensation by second filtering the reference data. Wherein a first performance of the first filtering may be different than a second performance of the second filtering.

Abstract: Various approaches for motion search refinement in a processing element are discussed. A k/2+L+k/2 register stores an expanded row of an L×L macro block. A k-tap filter horizontally interpolates over the expanded row generating horizontal interpolation results. A transpose storage unit stores the interpolated results generated by the k-tap filter for k/2+L+k/2 entries, wherein rows or columns of data may be read out of the transpose storage unit in pipelined register stages. A k-tap filter vertically interpolates over the pipelined register stages generating vertical interpolation results.

Abstract: Disclosed is a fine motion estimation device for high resolution including: a previous picture storage memory in which search area data of previous pictures for macroblocks of current pictures are stored; an FIR filter configured to perform FIR filtering on the search area data stored in the previous picture storage memory; a memory configured to differentiate and store the FIR filtered search area data; a QME data processing unit configured to generate reference area data for motion estimation in a ¼ pixel unit; a processing array unit configured to perform motion estimation in a ½ pixel unit and the macroblocks transmitted from the FIR filter and motion estimation in the ¼ pixel unit; and a control unit configured to control operations of the FIR filter, the processing array unit, and the QME data processing unit.

Abstract: Apparatus, systems and techniques based on an integer transform for encoding and decoding video or image signals, including transform of encoding and decoding of image and video signals and generation of an order-2N transform W from an order-N transform T in the field of image and video coding. For example, a retrieving unit is configured to retrieve an order-N transform T, where N is an integer; a deriving unit is configured to derive an order-2N transform W from the retrieved order-N transform T, and a transforming unit configured to generate an order-2N data Z using the derived transform W.

Abstract: In some embodiments, a method of determining encoding type and predictive mode(s) selections for a macroblock of a video frame is provided. In some embodiments, a general method 1) selects the encoding type (16×16 or 4×4) that is initially considered for a macroblock using an encoding type selection algorithm (based on an attribute of the macroblock that is easy to compute), 2) if the 16×16 encoding type is selected in step 1, consider the four 16×16 prediction modes that may be used on the macroblock using conventional methods or an improved 16×16 predictive mode search algorithm based on distortion thresholds, and 3) if the 4×4 encoding type is selected in step 1, select the 4×4 prediction mode to be used for each of the sixteen 4×4 blocks of the macroblock using conventional methods or an improved 4×4 predictive mode search algorithm based on the positional relationships between predictive modes.

Abstract: Methods of using motion estimation techniques with video encoders to provide significant data compression with respect to video signals so that the video signals may subsequently be reconstructed with minimal observable information loss. Methods include a fast fractional motion estimation scheme.

Abstract: In a method for the video coding of image sequences images in the image sequence are coded in a scaled manner, in such a way that the video data produced contains information which permits the images to be represented in a plurality of differing stages of image resolution, the latter being defined by the number of pixels per image representation. The coding is block-based, in such a way that to describe a displacement of parts of one of the images, said displacement being contained in the image sequence, at least one block structure that describes the displacement is created. Said block structure is configured from one block, which is subdivided into sub-blocks, whereby some of the sub-blocks are further subdivided into successively smaller sub-blocks. A first block structure is temporarily created for at least one first resolution stage and a second block structure is created for a second resolution stage, the first resolution stage having a lower number of pixels than the second resolution stage.

Abstract: Some embodiments facilitate encoding/decoding of a frame by organizing frame data in a storage structure in a novel manner. Specifically, in a portion of the storage structure allocated for a frame slice, used partition entries are stored in a first section of the allocated portion and unused partition entries are stored in a second section of the allocated portion, the first and second sections each comprising a continuous area of storage in the storage structure so that used partition entries are not interspersed with nonused partition entries. In some embodiments, additional data useful in the encoding or decoding of video data is determined and stored into the unused bytes of used partition entries (such as macroblock header data or canonical reference frame index data). In some embodiments, two or more identical partitions of a macroblock are coalesced into a single partition.

Abstract: A system and method for predicting an enhancement layer macroblock. A base layer frame is divided into intra-coded and inter-coded regions. If any portion of the enhancement layer macroblock is covered by both an intra-coded base layer macroblock and an inter-coded base layer macroblock, predictions utilizing the intra-coded and inter-coded macroblocks are established independently to generate at least two prediction values. The at least two prediction values are then combined to give a prediction from which the enhancement layer block is coded. Various embodiments serve to smooth the boundary effect between intra-coded regions and inter-coded regions inside the inter-layer prediction for extended spatial scalability.

Abstract: Sub-sampling pattern design for motion estimation in video compression. A motion estimation method divides a first frame into a plurality of macroblocks, performs block matching for a current macroblock pair on a candidate macroblock pair of a second frame. The current macroblock pair and the candidate macroblock pair are sampled according to a sub-sampling pattern, and an error measure between the current and candidate macroblock pairs is calculated. The sub-sampling pattern is constructed by a plurality of repeating units, and each repeating unit is composed of a first and a second pattern unit.

Abstract: A method includes projecting motion vectors describing a transformation from a previous video frame to a future video frame onto a plane between the previous video frame and the future video frame, detecting potential artifacts at the plane based on an intersection of a cover region and an uncover region on the plane, and analyzing a dissimilarity between a trial video frame and both the previous video frame and the future video frame. The trial video frame is generated between the previous video frame and the future video frame based on a frame rate conversion ratio derived from a source frame rate and a desired frame rate. The method also includes estimating reliability of the projected motion vectors based on the potential artifact detection and the dissimilarity analysis.

Abstract: A computer-implemented method for creating in the encoded domain one or more video frames from a compressed still video image wherein image content in the created video frames is translated in location by panning or scrolling or a combination thereof on a non-block basis (i.e. pixel-level). A new block formed from portions of two other blocks is created by processing the original two blocks with identity matrices based upon the shift amount. By performing the creation process in the encoded domain processing power requirements are reduced and image quality is increased.

Abstract: A method and apparatus utilizing a prediction guided decimated search motion estimation algorithm are provided. The prediction guided decimated search motion estimation algorithm generates a motion vector used to encode a macroblock in a frame from a video sequence. The algorithm includes generating full-pixel seed vectors, performing a full-pixel search around the generated seed vectors, which is followed by a fractional pixel search. The full-pixel seed vectors generated are a predicted motion vector and a hierarchical motion vector. A fractional pixel search may be conducted around a final motion vector generated by the full-pixel search and may include a half-pixel search and a quarter-pixel search. The prediction guided decimated search motion estimation algorithm can be implemented in both software and hardware. The algorithm is characterized by improved efficiency, scalability, and decreased complexity.

Abstract: A motion vector coding unit 117 executes processing including a neighboring block specification step (S100) of specifying a neighboring block which is located in the neighborhood of a current block; a judgment step (Steps S102, S104) of judging whether or not the neighboring block has been coded using a motion vector of another block; a prediction step (S106, S108) of deriving a predictive motion vector of the current block using a motion vector calculated from the motion vector of the other block as a motion vector of the neighboring block; and a coding step (S110) of coding the motion vector of the current block using the predictive motion vector.

Abstract: An encoding apparatus having a direct mode as a prediction mode, when a pixel with opposite parity from a top field to a bottom field, for example, is referred to for obtaining a reference vector in the direct mode, performs correction by adding or subtracting a value corresponding to a half pixel to or from a value of the obtained reference vector. The encoding apparatus, when a pixel with opposite parity is referred to for obtaining a first and a second direct vector by temporally scaling the corrected reference vector, performs correction by adding or subtracting a value corresponding to a half pixel to or from values of the obtained direct vectors.

Abstract: An object of the present invention is to express a predicted picture signal with light overheads, and to provide motion compensation of different pixel accuracy.

Abstract: A motion vector detecting device includes: a motion predicting and compensating circuit calculating cost values of a plurality of motion vectors, which is candidates of an optimal motion vector, using a cost function indicating an encoding efficiency with a first pixel precision every prediction mode and calculating the optimal motion vector with a second pixel precision and the cost value of the optimal motion vector with the second pixel precision using a profile of the cost values with the first pixel precision.

Abstract: Visibility of defects is improved for inspection of structures and the like, by generating an image having higher resolution than pixel resolution of a TV camera itself. An appearance inspection apparatus is provided with a TV camera; a camera driving device for making the TV camera scan an inspection object; an image capture device for capturing the image in the TV camera as a digital image; a camera motion measuring device for measuring scanning motion of the TV camera; a high definition image generating device, which generates a high definition image having a higher pixel resolution than that of the TV camera, based on the digital images captured by the image capture device and the TV camera scanning motion data measured by the camera motion measuring device; and a recording device which records and stores positional information of the inspection object.

Abstract: According to one embodiment, a method can generate an interpolation image signal including the integer-pixel and interpolation pixel values, based on, if the interpolation pixel is not located at the integer-pixel position and in a row of an integer-pixel, and the interpolation pixel is located at a first pixel position displaced from the integer-pixel in horizontal and vertical directions by a half pixel, applying a first filter, and if the interpolation pixel is located at a second pixel position displaced from the integer-pixel in the directions by a quarter pixel, applying a second filter, and if the interpolation pixel is not located at the first and second pixel positions, applying the second and first filters.

Abstract: A motion vector detecting device includes: a motion predicting and compensating means for calculating cost values of a plurality of motion vectors, which is candidates of an optimal motion vector, with a first pixel precision every prediction mode and calculating the optimal motion vector with a second pixel precision and a cost value of the optimal motion vector with the second pixel precision on the basis of a gradient of the cost values.

Abstract: An apparatus is described for attaching a motion search hardware assist unit to a processing element and its local memory. A current macro block storage unit is attached to a local memory interface unit for storage of a copy of a current macro block from the local memory. A search window reference storage unit having N rows is attached to a local memory interface unit for storage of a copy of N rows of pixels from a search window from the local memory. N independent arithmetic pipelines are attached to the current macro block storage unit and the search window reference storage. Each pipeline operates on one of the N rows of the search window reference storage unit and a corresponding row of the current macro block of the current macro block storage unit. An accumulator is attached to the N independent pipelines to accumulate results from the N arithmetic pipelines, to produce independent results for different organizations of macro blocks.

Abstract: A motion vector detection apparatus is configured to calculate a temporal distance between a frame to be coded and each of a plurality of reference candidate frames referred to by the frame to be coded. The motion vector detection apparatus searches for a candidate motion vector between the frame to be coded and each the plurality of reference candidate frames and detects a motion vector for the frame to be coded from the candidate motion vectors. In searching for and detecting a candidate motion vector, the motion vector detection apparatus changes an amount of the calculation performed during the detection of a candidate motion vector according to the calculated temporal distance between the frame to be coded and the reference candidate frame, and a coding type of the reference candidate frame.

Abstract: A system is configured to transcode a first MPEG stream to a second MPEG stream. The system includes a first MPEG decoder capable of decoding the first MPEG stream and a second MPEG encoder capable of producing the second MPEG stream. The second MPEG encoder is configured to maintain a decoded picture type of I, P, or B. The second MPEG encoder is also configured to maintain a decoded picture structure of frame or field, identify a metadata per each macroblock (MB) of an MB pair of the first MPEG stream, and determine whether to re-encode the MB into the second MPEG stream using one of a frame or a field mode based on the identified metadata. The second MPEG encoder is further configured to re-encode the MB pair into the second MPEG stream using one of the frame or the field mode based on the identified metadata.

Abstract: A method for interpolating a previous and subsequent image of an input image sequence, including: determining fidelity information for at least one motion vector that is descriptive for motion between a previous and subsequent image, wherein the fidelity information is descriptive for the level and accuracy of the motion, determining classification information for the at least one motion vector, wherein the classification information depends on the fidelity information, the classification information being descriptive for the motion type of the motion, and selecting an interpolation method in dependence of the determined classification information.

Abstract: A coding system includes an inter prediction block, a transform and quantization block, an encoding unit, and a reconstruction loop with an adaptive restoration block. The inter prediction block performs prediction on a current frame, the transform and quantization block performs transform and quantization processes on prediction residues, and the reconstruction loop reconstructs the current frame to generate the restored reconstructed samples. The adaptive restoration block performs restoration on processed data by considering a plurality of candidate restoration methods, selects one of the candidate restoration methods as a final restoration method, and generates adaptive restoration information corresponding to the final restoration method. The encoding unit encodes the prediction information and adaptive restoration information to generate an encoded bitstream.

Abstract: In some embodiments, a motion estimation search window cache is adaptively re-organized according to frame properties including a frame width and a number of reference frames corresponding to the current frame to be encoded/decoded. The cache reorganization may include an adaptive mapping of reference frame locations to search window cache allocation units (addresses). In some embodiments, a search window is shaped as a quasi-rectangle with truncated upper left and lower right corners, having a full-frame horizontal extent. A search range is defined in a central region of the search window, and is laterally bounded by the truncated corners.

Abstract: Embodiments of the invention are directed to a system and method for sub-pixel motion estimation for video encoding. The method includes providing a best match between a source frame and a reference frame by generating a plurality of non linear building surfaces, generating, in real time, an estimated matching criteria surface representing a matching criteria between the source frame and the reference frame based on the building surfaces and a plurality of sample points of an actual matching criteria surface and selecting, in real time, a position on the estimated matching criteria surface.

Abstract: A video encoder uses previously calculated motion information for inter frame coding to achieve faster computation speed for video compression. In a multi bit rate application, motion information produced by motion estimation for inter frame coding of a compressed video bit stream at one bit rate is passed on to a subsequent encoding of the video at a lower bit rate. The video encoder chooses to use the previously calculated motion information for inter frame coding at the lower bit rate if the video resolution is unchanged. A multi core motion information pre-calculation produces motion information prior to encoding by dividing motion estimation of each inter frame to separate CPU cores.

Abstract: Provided is a motion compensating apparatus that includes: a motion compensation position determining unit that determines, based on a motion vector, a position of pixels for which compensated pixels should be generated; a necessary pixel determining unit that determines pixels necessary for performing 6-tap filtering; a data transfer controlling unit that controls the order or the like of taking out data to be transferred; an intermediate pixel storage memory for storing pixel data with half-pixel accuracy; a high-order tap filtering unit that generates pixel data with half-pixel accuracy by successively performing filtering operations in a predetermined direction; and a linear interpolation calculating unit that performs linear interpolation based on the position of pixels to be motion compensated, and generates and outputs pixel data with motion compensation accuracy of less than half-pixel accuracy.

Abstract: A motion vector decoding apparatus decodes a coded motion vector of a current block in a moving picture. The motion vector decoding apparatus includes a neighboring block specification unit specifying a neighboring block which is located in the neighborhood of the current block and has already been decoded; a judgment unit judging whether or not the neighboring block has been coded using a motion vector of another block; a prediction unit deriving a predictive motion vector of the current block using a motion vector calculated from the motion vector of the other block as a motion vector of the neighboring block, when it is judged that the neighboring block has been coded using the motion vector of the other block; and a decoding unit decoding the coded motion vector of the current block using the predictive motion vector.

Abstract: A motion vector coding unit executes processing including a neighboring block specification step of specifying a neighboring block which is located in the neighborhood of a current block; a judgment step of judging whether or not the neighboring block has been coded using a motion vector of another block; a prediction step of deriving a predictive motion vector of the current block using a motion vector calculated from the motion vector of the other block as a motion vector of the neighboring block; and a coding step of coding the motion vector of the current block using the predictive motion vector.

Abstract: In video coding image sequences, images are coded in a scaled manner to obtain video data which represents the image in multiple steps from a defined image resolution or image quality (e.g., according to the data rate), and the resolution is defined by the number of image pixels of each represented image. Coding is block-based, i.e., for a description of an approximate movement of parts of one of the images in the image sequence, at least one block structure is produced describing the movement. The block structure produced includes a block divided into partial blocks which are in turn divided into sub-blocks. A first block structure is produced temporally for at least one first resolution level and a second block structure is produced for a second resolution level. The first resolution level has a lower image pixel number and/or image quality than the second resolution level.

Abstract: A motion vector coding unit 117 executes processing including a neighboring block specification step (S100) of specifying a neighboring block which is located in the neighborhood of a current block; a judgment step (Steps S102, S104) of judging whether or not the neighboring block has been coded using a motion vector of another block; a prediction step (S106, S108) of deriving a predictive motion vector of the current block using a motion vector calculated from the motion vector of the other block as a motion vector of the neighboring block; and a coding step (S110) of coding the motion vector of the current block using the predictive motion vector.

Abstract: Erasure information associated with a received group of encoded and interleaved data in a digital video broadcasting system is stored in a much compacted form. An erasure flag and an address of a last byte associated with the received group of encoded and interleaved data (a record) encapsulated in an MPE-FEC column will be stored in an erasure table. All bytes in the column preceding the last byte of the record will have the same erasure flag as the last byte. Erasure information deinterleaver 524 reads out the content of the erasure table (i.e., the erasure information) in a de-interleaving fashion; and the de-interleaved erasure information 525 are then applied with the de-interleaved coded signals 511 to an FEC decoder 526 to enhance the FEC decoding performance.

Abstract: It is possible to display a caption with an aspect ratio independent from the aspect ratio of a main video. When a flag indicating that the aspect ratio of the caption is 16:9 is set, the caption video image frame size (720×480) is converted so as to match the aspect ratio of 16:9 and the caption video obtained as the result is superimposed on the main video and displayed. That is, when the main video has an aspect ratio of 4:3, as shown in FIG. 19, reduction in the lateral direction is performed and the main video is displayed with addition of black tone at the right and left but the caption video is displayed with the aspect ratio of 16:9.

Abstract: Two-stage interpolation can be provided for frame prediction samples with quarter-pixel and finer accuracy. All samples of quarter-pixel and finer accuracy can use either half or full/integer-pixels in a bi-linear interpolation to allow for the use of higher accuracy motion vectors, such as one-eighth-pixel accuracy motion vectors. The motion vectors can be restricted in a manner such that they are not allowed to point to every possible sub-pixel sample on a sub-pixel grid, but rather a subset of those sub-pixel samples. In addition, the same full/integer and half-pixel samples that can be used to obtain a quarter-pixel sample can also be used to obtain a one-eighth-pixel sample that the quarter-pixel sample is connected to.

Abstract: Some embodiments provide an architecture for establishing multi-participant video conferences. This architecture has a central distributor that receives video images from two or more participants. From the received images, the central distributor generates composite images that the central distributor transmits back to the participants. Each composite image includes a set of sub images, where each sub image belongs to one participant. In some embodiments, the central distributor saves network bandwidth by removing each particular participant's image from the composite image that the central distributor sends to the particular participant. In some embodiments, images received from each participant are arranged in the composite in a non-interleaved manner. For instance, in some embodiments, the composite image includes at most one sub-image for each participant, and no two sub-images are interleaved.

Abstract: A method and apparatus for estimating motion of a pixel block in a first frame, the method including searching a first area in a second frame to identify a first matching block that corresponds to the pixel block, the first matching block including a first error value that is a minimum of at least one error criteria between the pixel block and the first matching block, calculating a first motion vector associated with the first matching block. The method further including searching a second area in the second frame to identify a second matching block that corresponds to the pixel block, the second matching block including a second error value that is a minimum of the at least one error criteria between the pixel block and the second matching block, calculating a second motion vector associated with the second matching block and selecting a final motion vector between the first and second motion vectors based on the first and second error value.

Abstract: A method for video encoding is disclosed. The method generally includes the steps of (A) generating first sub-pel data for at least one of (i) a motion estimation and (ii) a mode decision by first filtering reference data and (B) generating second sub-pel data for a motion compensation by second filtering the reference data. Wherein a first performance of the first filtering may be different than a second performance of the second filtering.

Abstract: A method and system for low-latency processing of intra-frame video pixel block prediction including: predicting a pixel block based on boundary pixels of left and upper neighbor blocks of said pixel block; subtracting said predicted pixel block from a source pixel block to generate a prediction error; forward transforming and quantizing said prediction error to generate a residual data; inverse transforming and quantizing said residual data; adding said predicted pixel block to said inverse transformed and quantized residual data to generate a reconstructed pixel block; pre-computing blocks of DC-coefficients used with luma and chroma intra prediction modes; pre-computing mode selection of a best prediction mode of said luma and chroma intra prediction modes; and outputting said residual data to be used in entropy or arithmetic coding, and a reconstructed data used for motion prediction.

Abstract: A method and system for interpolating video pixels is described, in which the values of a first quarter pixel, a half pixel and a second quarter pixel are calculated based on certain interpolation filter coefficients.

Abstract: A novel Lucas-Kanade sub-pixel motion estimation method is provided. The motion estimation algorithm enables the estimating of a motion vector with reduced computation cost while maintaining high sub-pixel accuracy. The novel algorithm consists of two processing stages. In the first stage, a conventional motion estimation method is applied to obtain the motion vector at integer-pixel level. In the second stage, the Lucas-Kanade algorithm is applied to improve the motion vector to sub-pixel accuracy based on gradient information. Experimental result shows that the proposed method reaches comparable PSNR performance as conventional ?-pel algorithm but with significant saving on computation cost.

Abstract: In one embodiment of the present application, a frame interpolation circuit performs frame interpolation on an input video signal, and outputs a drive video signal containing original and interpolation frames. A motion determination circuit outputs a control signal in accordance with the amount of motion, based on a motion vector obtained by a motion detection circuit. In accordance with the control signal, the frame interpolation circuit increases the proportion of interpolation frames contained in the drive video signal as motion in an image increases, while increasing the proportion of original frames contained in the drive video signal as the motion in the image decreases. As a result, any moving image blur due to following line of sight and noise generated in the interpolation frames are reduced.

Abstract: A method and apparatus for performing motion estimation in a digital video system is disclosed. Specifically, the present invention discloses a system that quickly calculates estimated motion vectors in a very efficient manner. In one embodiment, a first multiplicand is determined by multiplying a first display time difference between a first video picture and a second video picture by a power of two scale value. This step scales up a numerator for a ratio. Next, the system determines a scaled ratio by dividing that scaled numerator by a second first display time difference between said second video picture and a third video picture. The scaled ratio is then stored calculating motion vector estimations. By storing the scaled ratio, all the estimated motion vectors can be calculated quickly with good precision since the scaled ratio saves significant bits and reducing the scale is performed by simple shifts.

Abstract: Methods for motion estimation with adaptive motion accuracy of the present invention include several techniques for computing motion vectors of high pixel accuracy with a minor increase in computation. One technique uses fast-search strategies in sub-pixel space that smartly searches for the best motion vectors. An alternate technique estimates high-accurate motion vectors using different interpolation filters at different stages in order to reduce computational complexity. Yet another technique uses rate-distortion criteria that adapts according to the different motion accuracies to determine both the best motion vectors and the best motion accuracies. Still another technique uses a VLC table that is interpreted differently at different coding units, according to the associated motion vector accuracy.