Abstract: A video encoding/decoding apparatus according to the present invention acquires motion vector refinement information, performs motion compensation on the basis of a motion vector of a current block, refines the motion vector of the current block using at least one or both of the motion vector refinement information and the output of the motion compensation, and performs motion compensation using the refined motion vector.

Abstract: Disclosed herein are system, apparatus, article of manufacture, method and/or computer program product embodiments, and/or combinations and sub-combinations thereof, for recommending content based on user behavior tracking and analysis. Tracking data for a user indicating how the user interacted with a menu structure of a streaming media application is determined. A group from a plurality of user groups into which to group the user is selected based on the tracking data, wherein each group comprises data from a plurality of users. A plurality of content available for recommendation selection is determined. A first one of the plurality of content is selected as a content recommendation based on previous viewing patterns associated with the plurality of users of the selected group. The content recommendation is provided via the menu structure.

Abstract: A video encoding/decoding apparatus according to the present invention acquires motion vector refinement information, performs motion compensation on the basis of a motion vector of a current block, refines the motion vector of the current block using at least one or both of the motion vector refinement information and the output of the motion compensation, and performs motion compensation using the refined motion vector.

Abstract: Methods and apparatus are described for generating compelling preview clips of media presentations. Compelling clips are identified based on the extent to which human faces are shown and/or the loudness of the audio associated with the clips. One or more of these compelling clips are then provided to a client device for playback.

Abstract: An activity recognition system may comprise a local device and a server. The local device may be configured to: obtain a video; determine whether at least one human is present in one or more frames of the video; in response to determining the at least one human present in the video, determine if the at least one human in each of the frames corresponds to the same person; in response to determining that the at least one human in each of the frames corresponds to the same person, determine if the person is a stranger; and in response to determining that the person is not a stranger, transmit at least a portion of the video to the server. The server may be configured to: predict an activity class associated with the person; and trigger an alarm based on the activity class.

Abstract: An apparatus includes an interface and a processor. The interface may be configured to receive an array of scores. The processor may be configured to (i) parse the array of scores into a data flow including one or more operators, (ii) schedule the operators in one or more data paths, (iii) divide the array of scores into blocks of two or more scores by processing the array of scores using the one or more data paths, (iv) compute per-block information using the one or more data paths, and (v) compute local maxima position information for the array of scores by performing block based non-maximum suppression based on the per-block information using the one or more data paths. The data paths are generally implemented with a plurality of hardware circuits.

Abstract: Input video frames are encoded into a sequence of output video frames arranged in groups of pictures (GOP). Each input video frame is encoded into an output video frame using a respective compression value. Each GOP in the sequence of output video frames has a GOP length defined by a number of frames contained in the GOP. A level of motion and a level of light in a first input video frame are determined. If level of light is below a predetermined threshold value, encode the first input video frame using a predetermined constant GOP length and a compression value which is a decreasing function of the level of light. If level of light is above, encode the first input video frame using a GOP length which is a decreasing function of the level of motion, and a compression value which is a decreasing function of the level of light.

Abstract: A 3D model can be modeled using pattern-instance representation, wherein an instance component may be represented as transformation (for example, rotation, translation, and scaling) of a pattern. Quantization errors may be introduced when encoding rotation information, causing different vertex coordinate errors at different vertices of an instance. To efficiently compensate the vertex coordinate errors, an upper bound can be estimated for the vertex coordinate error of a vertex. Based on the upper bound, the codec decides whether the vertex coordinate error of the vertex needs to be compensated, and decides a quantization parameter for compensating the vertex coordinate error if compensation is needed. The upper bound can be estimated at both the encoder and decoder, and thus, no explicit signaling is needed to indicate whether vertex coordinate error compensation is used or to indicate the quantization parameter for the vertex coordinate error.

Abstract: A method is provided in one example embodiment and includes receiving video data at an adaptive bitrate (ABR) client that includes a buffer; determining whether a buffer level for the buffer is below a target buffer level; applying a random delay for a fetch interval associated with requesting the video data; and requesting a next segment of the video data after the random delay. The random delay can provide for a plurality of fetch times to become decorrelated from each other.

Abstract: Systems and methods for adaptive bitrate streaming of media including subtitles utilizing HTTP in accordance with embodiments of the invention are disclosed. One embodiment of the invention includes selecting a portion of the source media using the source encoder, transcoding the selected portion of the source media into a plurality of alternative portions of encoded video, where each alternative portion is encoded using a different set of encoding parameters and commences with an intra frame at the start of a closed group of pictures (GOP), writing each of the alternative portions of encoded video to a separate container file using the source encoder, writing a subtitle stream segment from each subtitle track corresponding to the selected portion of the source media to a separate container file using the source encoder, and associating a font file with at least one of the container files containing subtitle stream segments.

Abstract: A video call host module comprising a processor configured to decode video data corresponding to videos from endpoints and rank the videos based on motion indicators corresponding to each of the endpoints. The motion indicators are calculated from motion vectors corresponding to each of the videos. A predetermined number of highest-ranking videos are selected for display. A method of hosting a video call includes receiving encoded video data including motion vectors. Videos are ranked based on a motion indicator calculated from the motion vectors for each of the videos. Encoded video data is converted to decoded video data, and decoded video data corresponding to a predetermined number of the highest ranking videos is combined to create a single video.

Abstract: Methods and systems for providing a universal screen content codec are described. One method includes receiving screen content comprising a plurality of screen frames, wherein at least one of the screen frames includes a plurality of types of screen content. The method also includes encoding the at least one of the screen frames, including the plurality of types of screen content, using a single codec, to generate an encoded bitstream compliant with a standards-based codec. The plurality of types of screen content can include text, video, or image content. Blocks containing the various content types can be individually and collectively encoded.

Abstract: A sequence of images received from a camera is recorded according to a first procedure. The images are analyzed to identify events of interest, and upon identifying an event of interest, an area of interest of a plurality of said images is identified. The identified area is recorded according to a second procedure.

Abstract: A method and interactive system for the on-line transmission of a high-resolution video sequence composed of a succession of T images includes a step of selecting relevant images comprising at least the following steps: split each image to be transmitted at the instant t into a number N of zones, for each zone n determined in the previous step, calculate a value representative of the content of said zone, for each image to be transmitted, generate a vector representative of the content of said image containing the values obtained in the previous step, calculate a normalized coefficient of correlation ? between the reference vector determined for a previously selected image and that calculated for the current image, make a decision on the selection (or not) of the current image as a function of the value of the normalized correlation coefficient ?.

Abstract: Disclosed herein are methods and apparatuses for compressing a video signal. In one embodiment, the method includes storing a function derived from a set of human ratings in a memory, identifying within at least a portion of the video signal at least one content-based feature, inputting the at least one identified content-based feature into the stored function, determining a compression ratio based on the function using a processor and generating a compressed video signal at the determined compression ratio.

Abstract: A method for decoding video includes receiving a frame of the video that includes at least one slice and at least one tile. Each of the at least one slice and the at least one tile are not all aligned with one another. Each of the at least one slice is characterized that it is decoded independently of the other the at least one slice. Each of the at least one tile is characterized that it is a rectangular region of the frame and having coding units for the decoding arranged in a raster scan order. The at least one tile of the frame are collectively arranged in a raster scan order of the frame.

Abstract: The invention concerns a method which consists in analyzing a field of motion of images, estimated by using a first mesh, to detect a faulty area in the first mesh, and in locating a rupture line in said area; then generating a second mesh including a faultless part consisting of meshes of the first mesh outside the faulty area and two sub-meshes which overlap in a region including the rupture line. Each of the two sub-meshes includes respective meshes delimited by nodes including nodes shared with the faultless part, located at the boundary of the faulty area, and additional nodes not belonging to the faultless part, the rupture line being located between the respective nodes of the two sub-meshes shared with the faultless part. Said second mesh is used to finally estimate the field of motion in the group of images concerned.

Abstract: Aspects of a system for transforming uncompressed video traffic to network-aware Ethernet traffic with A/V bridging capabilities and A/V bridging extensions may include a graphics processing unit (GPU) that may enable encapsulation of uncompressed video data and/or auxiliary channel data within a Display Port protocol data unit (PDU). A LAN subsystem may enable encapsulation of the Display Port PDU within an encapsulating PDU. The LAN subsystem may enable determination of a traffic class designation associated with the encapsulating PDU. The LAN subsystem may enable transmission of the encapsulating PDU via a network based on the traffic class designation.

Abstract: An image encoding apparatus encodes image data and includes an image encoding unit that receives an input of the image data and image parameters and generates encoded image data by performing image encoding on the image data with reference to the image parameters and furthermore binarizing and arithmetically encoding the image data. A parameter processing unit outputs, as parameter information, parameters that are included in the image parameters and that are referred to when the encoded image data is arithmetically decoded, and encodes the image parameters to generate and output encoded image parameters. A stream generation unit generates a stream including the encoded image data obtained from the image encoding unit and the parameter information and the encoded image parameters that are outputted from the parameter processing unit.

Abstract: A method and apparatus for staggercasting includes encoding a first signal representing content having a time duration and a second signal also representing that content. A time delay period is specified for the time duration of the content. A composite signal, comprising the first and second encoded signals, is generated. In the composite signal the first encoded signal is delayed with respect to the second encoded signal by the time delay period for the time duration of the content. If an error is detected in the composite signal, then the received second encoded signal is decoded to produce the content, otherwise the delayed received first encoded signal is decoded to produce the content.

Abstract: A video transmission method is provided, which includes receiving state information from at least one mobile terminal that intends to perform a video stream service through a wireless network, determining a size of an image by selecting a specified spatial layer bit stream on the basis of the state information of the mobile terminal from a plurality of spatial layer bit streams generated at different bit rates during encoding of the bit stream, selecting a specified time and an SNR layer bit stream by increasing or decreasing time of the image and a layer position of the SNR layer bit stream on the basis of network parameters included in the state information of the mobile terminal, and transmitting the bit stream generated by extracting the specified layer bit stream of the selected layer to the mobile terminal.

Abstract: Rate control algorithms are adapted to cover multiple encoding standards. A rate controller includes an input adaptation interface, a core rate controller, and an output adaptation interface. The input adaptation interface converts rate control input parameters of a target encoding standard to corresponding native control input parameters of the rate controller. The core rate controller is coupled to the output of the input adaptation interface generating output parameters conforming to its native encoding standard. The output adaptation interface is coupled to the output of the core rate controller converting the generated output parameters to equivalent rate control output parameters conforming to the target encoding standard.

Abstract: A system for encoding a video stream into a processed video signal that includes at least one image, includes a pattern detection module for detecting a pattern of interest in the at least one image and identifying a region that contains the pattern of interest when the pattern of interest is detected. An encoder section, generates the processed video signal and wherein, when the pattern of interest is detected, a higher quantization is assigned to the region than to portions of the at least one image outside the region.

Abstract: Functionality is described for forming a summary representation of a video item to help a user decide whether to obtain a full version of the video item. The functionality operates by: (a) receiving a video item; (b) dividing the video item into a plurality of segments; (c) extracting at least one key frame from each of the plurality of segments to form a plurality of key frames; and (d) organizing the video segments into one or more groups corresponding to one or more respective scenes based on the plurality of key frames, to thereby form the representation of the video item. The functionality can be used to communicate search results to a user, to provide a sample of the video item in a message, etc.

Abstract: An apparatus and method for receiving multiple simultaneous stream bursts comprising determining if a frame size is smaller than a size of an available memory; determining if a priority of a first packet is lower than a priority of a second packet, wherein the frame size is of a frame for the second packet; requesting to abort processing the first packet and to de-assign the available memory from the first packet if the priority of the first packet is lower than the priority of the second packet; waiting for the size of the available memory to become equal or greater than the frame size if the priority of the first packet is not lower than the priority of the second packet; and assigning the available memory to the frame for the second packet.

Abstract: A computer implemented method of manipulating and displaying an MPEG stream is described. In one embodiment of the invention, a computer implemented method comprises defining a spatial location across a series of pictures of an MPEG stream; and for each picture of the series of pictures in the MPEG stream, partially decoding the picture to determine an area of the picture falling within the spatial location.

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, system, and medium are provided for formatting video frames such that a region of interest is emphasized and the video frames can be encoded, communicated, and rendered without excessive processing burdens. A region of interest is identified in a video frame and a feature mask is created that represents the region of interest. The feature mask can be used to crop the video frame to remove background images that are not within the region of interest and the cropped video frame can be overlayed on a simulated background before being encoded and communicated to a display device.

Abstract: Horizontal and vertical video shape padding (with or without attendant motion compensation) are facilitated through use of a two-dimensional small array of processing elements. Multiple pixel value assignments are determined and made with each clock cycle or iteration, thereby greatly reducing overall cycle time to complete the shape padding. This approach is compatible with MPEG 4.

Abstract: Systems and methods for efficiently reformatting video data in regions of video including occlusions are disclosed. In one embodiment, the method includes determining multiple motion vectors that link/relate matching blocks of two reference frames and calculating a measure related to overlap area of the matching blocks in one or both of the reference frames with a block to be constructed in an intermediate frame. The measure related to overlap area takes into account a particular interpolation phase of the frame being constructed in relation to the two reference frames. In one embodiment, a ranking of the measure related to overlap areas is used to classify the block to be constructed according to a degree of occlusion. In another aspect the location of the matching blocks in one or both of the reference frames is used in the classification of the block to be constructed.

Abstract: A data accessing and reproducing apparatus and method that permits a user to selected desired video clips and perform editing on the clips such that the time needed to display a thumbnail representation is minimized. In an edit mode, thumbnail pictures are arranged and displayed in the order of which edit results are reproduced. These thumbnails may be, for example, a lower resolution, proxy representation of the audio or video data. When data of thumbnail pictures are read from a disc, the data are read in the order of which they have been recorded. The order of which thumbnail pictures are displayed on the screen is different from the order of which they are arranged on the screen, and thus, the seek time for the disc becomes the shortest. Thus, the thumbnail pictures can be displayed at high speed.

Abstract: Learning-based image compression is described. In one implementation, an encoder possessing a first set of learned visual knowledge primitives excludes visual information from an image prior to compression. A decoder possessing an independently learned set of visual knowledge primitives synthesizes the excluded visual information into the image after decompression. The encoder and decoder are decoupled with respect to the information excluded at the encoder and the information synthesized at the decoder. This results in superior data compression since the information excluded at the encoder is dropped completely and not transferred to the decoder. Primitive visual elements synthesized at the decoder may be different than primitive visual elements dropped at the encoder, but the resulting reconstituted image is perceptually equivalent to the original image.

Abstract: A constant-quality rate control system and algorithm for multiple encoders with a single video source are disclosed. The system comprises a base encoder, a region of interest (ROI) detector, and at least one ROI encoder. The base encoder encodes a base video sequence which is a down-sampled version of a video sequence and sends the base video sequence to a multiplexer. The ROI detector extracts an ROI video sequence from the original video sequence. The ROI encoder encodes the ROI video sequence and sends it to the multiplexer. The ROI encoder uses information obtained during encoding of the base video sequence to improve the encoding of the ROI video sequence. The information of the base video sequence are used to better control the bit-rate and the quality of the ROI video sequence.

Abstract: Provided are methods and apparatuses for converting a lower bit-plane image, and methods and apparatuses for inverse-converting the lower bit-plane image, which increase the correlation of the lower bit-plane image. According to the method of converting a lower bit-plane image, pixels of the lower bit-plane image are sequentially added or subtracted so as to generate a converted lower bit-plane image. Accordingly, the lower bit-plane image has higher regularity, and thus the efficiency of image processing using the lower bit-plane image is improved.

Abstract: The invention discloses an image processing apparatus for improving the definition of an input image. The input image consists of a plurality of pixels, and the pixels are divided into an edge portion and a background portion. The image processing apparatus includes a compound operating unit, a sharpness processing unit, and an edge detecting unit. The compound operating unit is used for generating a noise-eliminated image by subjecting the input image to an open-close process and a close-open process. The sharpness processing unit is used for generating an enhanced image by subjecting the input image to a sharpness process. The edge detecting unit is used for receiving the noise-eliminated image and the enhanced image and determining whether each pixel of the input image belongs to the edge portion or the background portion, so as to selectively output the gray level of each pixel of the enhanced or noise-eliminated image.

Abstract: Real-time image stabilization using computationally efficient corner detection and correspondence is provided. The real-time image stabilization performs a scene learning process on a first frame of an input video to obtain reference features and a detection threshold value. The presence of jitter is determined in a current frame of the input video by comparing features of the current frame against the reference features using the detection threshold value. If the current frame is found to be unstable, corner points are obtained from the current frame. The obtained corner points are matched against reference corner points of the reference features. If the number of matched corner points is not less than a match point threshold value, the current frame is modeled using random sample consensus. The current frame is corrected to compensate for the jitter based on the results of the modeling.

Abstract: An apparatus and method for displaying topography around a position of a mobile object such as a vehicle in 3-dimensional graphics. The method includes predicting a candidate region for the mobile object's position at a display time after a predetermined period of time using current position and motion information of the mobile object; loading 3-dimensional graphic data corresponding to the predicted candidate region to memory; and rendering and displaying data corresponding to a region around a position of the mobile object at the display time among the 3-dimensional graphic data loaded to the memory.

Abstract: The invention relates to motion estimation and compensation between two given images represented by a polygonal mesh. It provides a solution to the problems of crowding and uncovering of mesh units when the mesh is shifted. More specifically, the invention relates to a method for the encoding of a mesh that represents an image of a moving sequence. According to the invention, such a mesh is larger in size than the image.

Abstract: Techniques for performing mesh-based video compression/decompression with domain transformation are described. A video encoder partitions an image into meshes of pixels, processes the meshes of pixels to obtain blocks of prediction errors, and codes the blocks of prediction errors to generate coded data for the image. The meshes may have arbitrary polygonal shapes and the blocks may have a predetermined shape, e.g., square. The video encoder may process the meshes of pixels to obtain meshes of prediction errors and may then transform the meshes of prediction errors to the blocks of prediction errors. Alternatively, the video encoder may transform the meshes of pixels to blocks of pixels and may then process the blocks of pixels to obtain the blocks of prediction errors. The video encoder may also perform mesh-based motion estimation to determine reference meshes used to generate the prediction errors.

Abstract: An object-region-data describing method for describing object region data relating to a time-series variation of an object region in video data including a plurality of frames, the method comprising obtaining a conversion parameter representing conversion from a reference object region into a target object, approximating a time-series variation of the conversion parameter by an approximate function, and describing the object region data using an approximate function parameter identifying the approximate function and information on the reference object region. Thus, it is possible to describe a region of the desired object in video data by a small quantity of data and facilitate generating the object regions and handling data on the object region.

Abstract: Briefly, in accordance with one embodiment, a technique for coding the movement of a head or face from a sequence of images is disclosed. A variety of potential alternative embodiments are discussed.

Abstract: The invention relates to a method of coding segmented pictures, or partitions, divided into homogeneous regions. Said method, called intra-mode coding process, comprises, for each successive partition, the steps of translating the picture of the labels associated to the regions into a description in terms of a contour element chain in which the elements are defined by means of their movements through successive basic cells, tracking inside each successive cell each contour from its initial contour point, previously extracted, to its end by storing chain symbols corresponding both to input, internal and output contour elements of said cell and to priorities between possible multiple outputs elements, and repeating these steps up to the end of each successive contour segment of the concerned partition.

Abstract: As a unit element for defining a describing procedure for description of a feature data characteristic of the content of a video data, a descriptor is formed in a description format structured to have its attribute as lower-level element and which may include other descriptor as attribute. That is, the description format for the descriptor consists of a feature indicative of a descriptor or attribute as lower level element, and a feature type indicative of the data type of the feature.

Abstract: The invention relates to scalable video encoders, and, more precisely, to an encoder comprising a base layer and at least an enhancement layer circuit, said circuits being capable of generating enhancement layer video data associated with compressed base layer video data and suitable for transmission to a video decoder. In said encoder, the implemented encoding method, based on a hierarchical triangular mesh representation to which a matching pursuit error coding step is specifically adapted, comprises a base layer encoding step and enhancement layer encoding steps. The invention also relates to a corresponding decoding method.

Abstract: The invention relates to a method for estimating the motion between two digital images and to the use of this method for coding images. The aim of the method is to eliminate the disturbances caused by the mesh being overturned during the motion estimation process. According to the invention, the method consists of carrying out a first motion estimation in order to identify the faulty mesh, i.e. the mesh that is overturning or is overlapping after the displacement vectors have been applied; then carrying out a second motion estimation, excluding at least said faulty mesh, in order to minimize prediction error on the remaining mesh of the model.

Abstract: The present invention relates to a method of encoding, which receives an input video signal (IS) for providing a compressed video data signal (CS). Said method comprises the steps of segmentation (SEG) of the input video signal for extracting the video object planes, hereafter referred to as VOP, contained in the video scene; detection (DET) of a non-visible VOP; correction (COR) for converting a bidirectionally predictive coded VOP preceding, in a display order, a non-visible intra or predictive coded VOP that has been detected (Y) by the detection step, into an intra or predictive coded VOP; encoding (ENC) the converted VOP, if a non-visible VOP has been detected by the detection step, or the regular VOP in the opposite case (N), resulting in the compressed video data signal. Use: MPEG-4 encoder.

Abstract: The present invention includes a method and system for image compression using block size heuristics. A method for motion searching a video frame is disclosed including iteratively decreasing block size until a rate-distortion (RD) has been minimized. A method for compressing motion video images is disclosed. Additionally, a system for transmitting and receiving video images is disclosed. The system may be a video conferencing system.

Abstract: An apparatus may include a sort circuit for receiving first and second baseband signals, where the sort circuit shifting frequencies of the first and second baseband signals. The apparatus may also include a removal circuit for receiving the shifted first and second baseband signals and for combining the shifted first and second baseband signals to provide a frequency-modulated signal, and a symbol timing restoration circuit for measuring a timing error in related symbols of the frequency-modulated signal, for generating an address selection signal that is proportional to the timing error, in response to a carrier restoration signal, and for indicating restoration of the carrier.

Abstract: There is provided an image failure detection unit in redundant duplex transmission in which a failure occurring on a regular link is detected at in real time, and the link is instantaneously switched to a normal backup link, thereby preventing any failure from occurring on an output image or reducing the failure. Units for calculation of the image features calculate image features of links A and B, respectively. A comparison unit compares the image features, and then, it is judged that both of the links are normal if there is not difference between the image features. To the contrary, if there is a difference, it is judged that there is a probability of occurrence of a failure on one of the links. Thereafter, a normal/corrupted information memory stores a small region relating to the image features. Units for detection of image feature differences determine image feature differences for the links, respectively.

Abstract: The invention relates to motion estimation and compensation between two given images represented by a polygonal mesh. It provides a solution to the problems of crowding and uncovering of mesh units when the mesh is shifted.