Abstract: The invention relates to a method for automatic detection and tracking of one or multiple objects of interest in a video sequence comprising several successive frames (1) captured with a moving camera, comprising:—segmenting (32) each frame into regions depending on color and producing a segmented image (1?);—deducing (34) from each segmented image (1?) a region adjacency graph (11) comprising nodes and edges, wherein each node defines a region and each edge links two nodes of adjacent regions; wherein the region adjacency graphs (11) form a set of region adjacency graphs;—extracting (36) from the set of region adjacency graphs frequent spatio-temporal patterns (13);—determining (38) degrees of dissimilarity between the trajectories of the extracted frequent spatio-temporal patterns (13) between successive frames (1); and—clustering (39) the frequent spatio-temporal patterns (13) to produce at least one cluster (25, 26, 27) corresponding to one object of interest among said one or multiple objects of interest

Abstract: Methods and apparatuses for performing arithmetic operations efficiently and quickly are described. Such arithmetic operations include, but are not limited to, multiplying 2N bit integers, multiplying multiple N-bit integers simultaneously, multiplying 2N bit complex numbers, and other multiplication operations involving coefficients, complex numbers, and complex conjugate numbers.

Abstract: A picture coding apparatus includes a motion vector estimation unit and a motion compensation unit. The motion vector estimation unit selects one method for deriving a motion vector of a block to be motion-compensated, depending on a motion vector of a block located in a corner of a decoded macroblock from among a group of blocks that compose the decoded macroblock corresponding to the current macroblock to be coded and determines the motion vector derived by the selected method for derivation to be a candidate of the motion vector of the current macroblock to be coded. The motion compensation unit generates a predictive image of the block to be motion-compensated based on the estimated motion vector.

Abstract: A video surveillance system is set up, calibrated, tasked, and operated. The system extracts video primitives and extracts event occurrences from the video primitives using event discriminators. The system can undertake a response, such as an alarm, based on extracted event occurrences.

Abstract: A method for temporal filtering based on motion detection between non-adjacent pictures is disclosed. Step (A) of the method may compute a plurality of motion scores by motion detection between a target picture in a sequence of pictures and a non-adjacent reference picture in the sequence of pictures. Step (B) may temporal filter the target picture with an adjacent reference picture in the sequence of pictures based on the motion scores to generate a filtered picture. At least one of (i) the motion scores and (ii) the generation of the filtered picture is controlled by one or more gain settings in a circuit.

Abstract: Interpolated video tagging techniques are described. An example method includes identifying, a sequence of video frames included in a plurality of video frames of a video, where an object exists in at least one video frame of the sequence of video frames, determining an interpolation function associated with the existence of the object in the sequence of video frames, where the interpolation function specifies a time domain of the video in which the interpolation function is valid, and transmitting the video, an indication of the object, and an indication of the interpolation function.

Abstract: Methods and apparatus are provided for adaptive coupled pre-processing and post-processing filters for video encoding and decoding. The apparatus and method encode input data for a picture into a resultant bitstream, wherein said video encoder comprises a pre-filter and a post-filter coupled to the pre-filter, wherein said pre-filter filters the input data for the picture and the post-filter filters in-loop reconstructed data for the picture.

Abstract: A method for processing image data representing a three-dimensional volume, the data comprising image values for a three-dimensional grid of voxels, comprising: starting from a given voxel building up a vector path along a first dimension of the three dimensional volume, connecting a number of voxels of two-dimensional slices of the grid adjacent in the first dimension, the connected voxels representing a spatial neighborhood of the given voxel, including structurally related voxels of the three-dimensional grid; averaging the image values of the voxels of the vector path to obtain a first averaged value, assigned to the given voxel position; repeating the aforementioned steps for a number of voxels; repeating the aforementioned steps on the first averaged values, employing vector paths along a second dimension being different from the first dimension, obtaining second averaged values; and on the second averaged values repeating the aforementioned steps, employing vector paths along a third dimension being di

Abstract: A component for controlling playback of digital media objects using a single control input receives, from a server, a plurality of feature vectors, each feature vector representing one of a plurality of media objects, and receives one or more skip requests from the single control input. In response to a first skip request, the component selects a feature vector from among the plurality of feature vectors and instructs a media object playback device to playback a media object represented by the selected feature vector. In response to a second or successive skip request, the component selects another feature vector from among the plurality of feature vectors that is a distance away from the previously selected feature vector and instructs the media object playback device to playback another media object represented by the selected other feature vector.

Abstract: Provided is an image coding method that partitions an input image signal into processing units, and that codes the partitioned image to generate a code sequence. In particular, the image coding method determines a partitioning pattern for hierarchically partitioning the input image signal in order starting from a largest unit of the processing units, generates partition information indicative of the partitioning pattern, and codes partition information. The partition information includes maximum used hierarchy depth information indicative of a maximum used hierarchy depth which is a hierarchy depth of a deepest processing unit of the processing units included in the partitioning pattern.

Abstract: Methods and apparatuses for data pruning for video compression using example-based super resolution are provided. A method and apparatus for encoding is provided in which patches of video are extracted from input video, grouped together using a clustering method, and representative patches are packed into patch frames. The original video is downsized and sent either along with, or in addition to, the patch frames. At a decoder, the method and apparatus provided extract patches from the patch frames and create a patch library. The regular video frames are upsized and the low resolution patches are replaced by patches from the patch library by searching the library using the patches in the decoded regular frames as keywords. If there are no appropriate patches, no replacement is made. A post processing procedure is used to enhance the spatiotemporal smoothness of the recovered video.

Abstract: The invention introduces a method for determining frame resolution, performed by a processing unit, which contains at least the following steps: obtaining a YUV video file, which contains a video stream; selecting a plurality of frame resolutions from a frame resolution table as frame resolution candidates according to the length of the video stream; repeatedly executing a loop m times, and in each iteration, obtaining at least two frames of the video stream according to one unexamined frame resolution candidate and calculating the correlation between the obtained frames, wherein m represents the quantity of frame resolution candidates; determining the best frame resolution from the frame resolution candidates according to the calculated correlations; and outputting the best frame resolution, thereby enabling a video decoder to segment the video stream into frames according to the best frame resolution and decode the segmented frames.

Abstract: A method for encoding pixels in an image, includes: encoding an offset to be applied to a reference frame; generating predicted pixels of the pixels in the image, based on interpolating pixels in the reference frame and then adding the offset to the interpolated pixels; and encoding residual signals that are differences between the pixels in the image and the predicted pixels, wherein the encoding of the residual signals comprises: skipping both transform and quantization with respect to residual signals of which an encoding scheme corresponds to skipping of both transform and quantization, the encoding scheme related to transform and quantization, skipping transform and performing quantization with respect to residual signals of which the encoding scheme corresponds to skipping of transform, and performing transform and quantization with respect to residual signals of which the encoding scheme corresponds to skipping neither transform nor quantization.

Abstract: An apparatus for generating a reference frame is disclosed including: an interpolation unit for interpolating a reference frame per fractional pixel; a virtual encoder for determining a block mode of a current frame and a motion vector according to the block mode by using an interpolated reference frame; an offset calculator for calculating an offset for each pixel of the interpolated reference frame by using the block mode and the motion vector; and an offset adder for generating an offset interpolated reference frame by adding the offset for each pixel to each pixel of the interpolated reference frame. By adding the offset calculated according to the block mode to the reference frame for the inter prediction and making the reference frame to be used for the motion estimation and compensation similar to the current frame to be encoded, motion prediction is accurate and compression efficiency is improved.

Abstract: A video encoding method is provided. The method includes: encoding a current region of a video by performing a transformation on the current region by using transformation units in a variable tree-structure which are determined from among transformation units that are hierarchically split from a base transformation unit with respect to the current region and which are generated based on a maximum split level of a transformation unit. The method further includes outputting encoded data of the current region, information about an encoding mode, and transformation-unit hierarchical-structure information including maximum size information and minimum size information of the transformation unit with respect to the video.

Abstract: A method for selecting a motion vector predictor is provided. The method of selecting a motion vector predictor includes the steps of selecting motion vector predictor candidates for a current block and selecting a motion vector predictor of the current block out of the motion vector predictor candidates, wherein the motion vector predictor candidates for the current block include a motion vector of a first candidate block which is first searched for as an available block out of left neighboring blocks of the current block and a motion vector of a second candidate block which is first searched for as an available block out of upper neighboring blocks of the current block.

Abstract: A method may include determining movement between a first frame of a monoscopic video and a second frame of the monoscopic video. The method may also include analyzing a camera effect based on the movement, generating a first modified frame based on the movement and the camera effect, and generating a second modified frame based on the movement and the camera effect. The method may also include generating a left-eye viewing frame of a stereoscopic video based on the camera effect and at least one of the first modified frame and the second modified frame. The method may further include generating a right-eye viewing frame of the stereoscopic video based on the camera effect and at least one of the first modified frame and the second modified frame.

Abstract: A method for encoding an interlaced video comprising a sequence of alternating TOP and BOT fields is disclosed. The method for encoding comprises: producing a sequence of frames, each frame being determined from at least two consecutive fields, each field forming a rectangular region of the frame, encoding each frame of the sequence to produce an encoded video of the sequence of frames.

Abstract: The invention provides a video codec. In one embodiment, the video codec is coupled to an outer memory storing a reference frame, and comprises an interface circuit, an in-chip memory, a motion estimation circuit, and a controller. The interface circuit obtains in-chip data from the reference frame stored in the outer memory. The in-chip memory stores the in-chip data. The motion estimation circuit retrieves search window data from the in-chip data with a search window, and performs a motion estimation process on a current macroblock according to the search-window data. The controller shifts the location of the search window when the current macroblock is shifted, marks a macroblock shifted out from the search window as an empty macroblock, and controls the interface circuit to obtain an updated macroblock for replacing the empty macroblock in the in-chip memory from the reference frame stored in the outer memory.

Abstract: A video decoding method for performing video decoding, in which a maximum split level of a transformation unit is determined based on maximum split information for an intra mode when information about the prediction mode indicates that a coding unit is predicted according to the intra mode. Accordingly, the information about the prediction mode indicates that the coding unit is predicted according to the intra mode, rather than according to any prediction modes.

Abstract: The present disclosure provides an image encoding/decoding apparatus and method, and a reference picture indexing apparatus and method. The reference picture indexing apparatus includes: a candidate reference picture setting unit for setting candidate reference pictures; a priority order determining unit for determining a priority order of the candidate reference pictures set by the candidate reference picture setting unit, according to a current block to be encoded; and a reference picture indexing unit for adaptively indexing a reference picture according to the current block based on the priority order determined by the priority order determining unit.

Abstract: A method of generating a stereoscopic video may include determining movement between a first frame of a monoscopic video and a second frame of the monoscopic video. The method may further include analyzing a camera effect based on the movement. Additionally, the method may include generating a left-eye viewing frame of a stereoscopic video based on the camera effect analysis and generating a right-eye viewing frame of the stereoscopic video based on the camera effect analysis.

Abstract: This invention is a digital signal processor form plural sums of absolute values (SAD) in a single operation. An operational unit performing a sum of absolute value operation comprising two sets of a plurality of rows, each row producing a SAD output. Plural absolute value difference units receive corresponding packed candidate pixel data and packed reference pixel data. A row summer sums the output of the absolute value difference units in the row. The candidate pixels are offset relative to the reference pixels by one pixel for each succeeding row in a set of rows. The two sets of rows operate on opposite halves of the candidate pixels packed within an instruction specified operand. The SAD operations can be performed on differing data widths employing carry chain control in the absolute difference unit and the row summers.

Abstract: In a case where a weighted mean value is calculated by calculating weights according to a degree of similarity based on a target pixel, the mean value is affected by noise included in the target pixel. The mean value thereby converges at a value different from a true pixel value without noise and this appears as a remaining noise. To solve this problem, an output value of the target pixel is determined based on a correlation established among a weighted mean value of multiple reference pixels, a pixel value of the target pixel, and the true value of the target pixel.

Abstract: An image encoding device that appropriately estimates a prediction vector by avoiding an uncoded macro block (MB) is disclosed. The device determines a start point of searching a motion vector for an encoding target region, from the prediction vector for the target region which is determined with motion vectors for surrounding blocks around the target region and in accordance with a predetermined rule, and starts searching for the motion vector from the search start point. If the size of the encoding target region cannot be acquired during calculation of the prediction vector, the device determines the size of the encoding target region, from the size of an encoded region other than the target region. If at least one of the motion vectors for the surrounding blocks cannot be acquired, the device calculates the prediction vector using alternatively a motion vector of an encoded block outside of the target region.

Abstract: A motion information storing method and a TMVP deriving method using the motion information storing method are provided. The motion information storing method includes a step of determining a representative block out of blocks of a motion data storage unit in a co-located picture with a current picture and a step of storing motion information of the representative block as motion information of the motion data storage unit, wherein the motion information of the representative block is used as motion information of the blocks included in the motion data storage unit.

Abstract: A method of processing image data includes detecting a contour between an occlusion region and a non-occlusion region in a motion vector field of input video, generating a first contour and a second contour from the detected contour based on motion vectors of blocks corresponding to the detected contour, and separating the occlusion region into a first occlusion region adjacent to the first contour and a second occlusion region adjacent to the second contour. The method further includes reconstructing motion vectors of the first occlusion region based on motion vectors of the first contour and reconstructing motion vectors of the second occlusion region based on motion vectors of the second contour.

Abstract: A method for selecting a motion vector predictor is provided. The method of selecting a motion vector predictor includes the steps of selecting motion vector predictor candidates for a current block and selecting a motion vector predictor of the current block out of the motion vector predictor candidates, wherein the motion vector predictor candidates for the current block include a motion vector of a first candidate block which is first searched for as an available block out of left neighboring blocks of the current block and a motion vector of a second candidate block which is first searched for as an available block out of upper neighboring blocks of the current block.

Abstract: A method for selecting a motion vector predictor is provided. The method of selecting a motion vector predictor includes the steps of selecting motion vector predictor candidates for a current block and selecting a motion vector predictor of the current block out of the motion vector predictor candidates, wherein the motion vector predictor candidates for the current block include a motion vector of a first candidate block which is first searched for as an available block out of left neighboring blocks of the current block and a motion vector of a second candidate block which is first searched for as an available block out of upper neighboring blocks of the current block.

Abstract: A computer implemented method is provided for computing a two-way optical flow between a reference frame and one or more associated frames in an optical flow module. A forward warping operator and a backward warping operator can be generated between the reference frame and each of the one or more associated frames in a warping module. The forward warping operator and the backward warping operator provide motion compensation between the frames. Weights for each of the forward warping operators and the backward warping operators can be computed in a weight module. The weights correspond to uncertainty of motion estimation. A super resolution iteration algorithm can be calculated in a super-resolution iteration module.

Abstract: A video display apparatus includes a video display unit configured to display a video image, a viewer detection unit configured to detect a position of a viewer, a moving speed measurement unit configured to measure a moving speed of the viewer, a display area determination unit configured to determine, based on the position and moving speed of the viewer, a video display area where the video image is displayed on the video display unit, and a video control unit configured to display the video image in the video display area determined by the display area determination unit.

Abstract: An image search method may include: determining a quadrant of a predicted motion vector; calculating a tilt value of a first reference frame and a tilt value of a second reference frame using the predicted motion vector; deciding a search area for uneven hexagon search in response to the quadrant of the predicted motion vector and the calculated tilt values; performing the uneven hexagon search with respect to the decided search area; and/or comparing a result of the performed uneven hexagon search with a threshold value to determine termination of the uneven hexagon search. The second reference frame is earlier-in-time relative to the first reference frame.

Abstract: A method of scalable video coding in which the video signal is processed and coded at various layers of spatial resolution where between each pair of spatial layers a prediction mechanism is enabled which allows a comparison at homologous layers of spatial resolution with the aim to obtain a good efficiency.

Abstract: A motion-vector-setting section (31) sets a motion vector in units of pixel in a target image. Based on the motion vector, a target-pixel-setting section (35) sets a target pixel for each image in plural images to be processed. A motion-blur-amount-setting section (33) sets a motion blur amount in units of pixel based on the motion vector and the exposure-time ratio set in units of image in the exposure-time-ratio-setting section (32). A processing-region-setting section (36) sets processing regions corresponding to the target pixel for each of the plural images based on the motion blur amount. A processing-coefficient-setting section (37) sets processing coefficients based on the motion blur amount.

Abstract: A method for measuring motion at a horizontal and vertical position between video fields of opposite parity comprising the steps of measuring the signal values of at least two vertically adjacent pixels from a video field of one parity and at least two vertically adjacent pixels from a video field of the opposite parity such that when taken together, the pixels represent contiguous samples of an image at said horizontal and vertical position, and determining whether the signal value of any of the pixels lies between the signal values of adjacent pixels in the field of opposite parity and in response outputting a zero motion value, otherwise, outputting a motion value equal to the lowest absolute difference between any of the pixels and its closest adjacent pixel in the field of opposite parity.