Abstract: A method of point cloud geometry encoding includes receiving a slice of a point cloud frame for encoding, and constructing an octree representing a geometry of points in a bounding box of the slice where a current node of the octree is partitioned with a quadtree (QT) partition or a binary tree (BT) partition. The constructing includes determining a value of a partitionSkip variable specifying a partition type and a partition direction of the current node of the octree.

Abstract: A method and a system for training a machine-learning algorithm (MLA) to determine a user class of a user of an electronic device are provided. The method comprises: receiving a training audio signal representative of a training user utterance; soliciting, by the processor, a plurality of assessor-generated labels for the training audio signal, the given one of the plurality of assessor-generated labels being indicative of whether the training user is perceived to be one of a first class and a second class; generating an amalgamated assessor-generated label for the training audio signal, the amalgamated assessor-generated label being indicative of a label distribution of the plurality of assessor-generated labels between the first class and the second class; generating a training set of data including the training audio signal and the amalgamated assessor-generated to train the MLA to determine the user class of the user producing an in-use user utterance.

Abstract: Coding schemes for coding a spatially sampled information signal using sub-division and coding schemes for coding a sub-division or a multitree structure are described, wherein representative embodiments relate to picture and/or video coding applications.

Abstract: This disclosure describes entropy coding techniques for media data coded using neural-based techniques. A media coder is configured to determine a probability distribution function parameter for a data element of a data stream coded by a neural-based media compression technique, wherein the probability distribution function parameter is a function of a standard deviation of a probability distribution function of the data stream, determine a code vector based on the probability distribution function parameter, and entropy code the data element using the code vector.

Abstract: Artificial image generation may include obtaining a source image, identifying quantization information from the source image, wherein identifying the quantization information includes identifying multiresolution quantization interval information from the source image, generating a restoration filtered image by restoration filtering the source image, generating a constrained restoration filtered image by constraining the restoration filtered image based on the quantization information, obtaining an unconstrained artificial image based on the constrained restoration filtered image and a generative artificial neural network obtained using a generative adversarial network, obtaining the artificial image by constraining the unconstrained artificial image based on the quantization information, and outputting the artificial image.

Abstract: Disclosed herein is a method for transmitting point cloud data, the method including encoding point cloud data, and transmitting a bitstream containing the point cloud data. Disclosed herein is a method for receiving point cloud data, the method including receiving a bitstream containing point cloud data, decoding the point cloud data, and rendering the point cloud data.

Abstract: Aspects of the disclosure provide a method and an apparatus for video coding. A processing circuitry determines a process unit size for a decoder-side motion vector refinement (DMVR). Then, the processing circuitry enables an application of the DMVR on a current block when a size of the current block is smaller than or equal to the process unit size; and disables the application of the DMVR on the current block when the size of the current block is larger than the process unit size. In an example, the application of the DMVR refines the motion vectors with or without a template block.

Abstract: A method for deblurring a target image includes receiving a burst of images that includes the target image; partitioning respective images of the burst of images into respective patches; converting, to a frequency domain, the respective patches into respective transform patches; selecting a first set of corresponding transform patches from the respective transform patches, where the first set of the corresponding transform patches includes a respective transform patch for a respective image of the burst of images; obtaining, using a neural network, respective weight maps for the corresponding transform patches; obtaining a deblurred transform patch by combining the first set of corresponding transform patches using the respective weight maps; obtaining a first deblurred patch by converting the deblurred transform patch to a pixel domain; and obtaining a deblurred image of the target image using the first deblurred patch.

Abstract: Disclosed herein is a method for transmitting point cloud data. The transmission method may include encoding point cloud data, and transmitting the point cloud data. Disclosed herein is a method for receiving point cloud data. The reception method may include receiving point cloud data, decoding the point cloud data, and rendering the point cloud data.

Abstract: The present disclosure provides a method and apparatus for subband-coding a frequency conversion unit and a video encoding/decoding method a video encoding/decoding apparatus using the same. The method and the apparatus for subband-coding the frequency conversion unit include generating an encoding stream by splitting the frequency conversion unit into one or more frequency domains, generating frequency domain encoding information according to whether there is a non-zero frequency coefficient in each frequency domain, scanning the frequency coefficient of each frequency domain to generate a frequency domain frequency coefficient stream, and binarizing and encoding the frequency domain encoding information and the scanned frequency domain frequency coefficient stream.

Abstract: Provided are a method and apparatus, which, during video encoding and decoding processes, obtain chroma intra prediction mode information about a current chroma block, when the chroma intra prediction mode information indicates a direct mode (DM), determine a luma block including a luma sample corresponding to a chroma sample at a lower-right location with respect to a center of the current chroma block, determine a chroma intra prediction mode of the current chroma block based on an intra prediction mode of the determined luma block, and perform intra prediction on the current chroma block, based on the determined chroma intra prediction mode.

Abstract: Embodiments of systems and methods for intelligently selecting audio signatures based upon context information to perform management actions are described. In some embodiments, an Information Handling System (IHS) may include a processor and a memory coupled to the processor, the memory having program instructions stored thereon that, upon execution, cause the IHS to: select, based upon context information, a subset of a plurality of audio signatures, compare a received audio input to at least one audio signature among the subset of audio signatures to the exclusion of any other audio signature of the plurality of audio signatures, and, in response to the comparison indicating a match, perform one or more management actions.

Abstract: A computer-implemented method, system and computer program product for compressing video. A set of video frames is partitioned into two subsets of different types of frames, a first type and a second type. The first type of frames of videos is compressed to generate a first representation by a first stage encoder. The first representation is then decoded to reconstruct the first type of frames using a first stage decoder. The second type of frames of video is compressed to generate a second representation that only contains soft edge information by a second stage encoder. A generative model corresponding to a second stage decoder is then trained using the first representation and the reconstructed first type of frames by using a discriminator employed by a machine learning system. After training the generative model, it generates reconstructed first and second types of frames using the soft edge information.

Abstract: Disclosed herein is a transmitting method and a receiving method of point cloud data. The transmitting method may include encoding point cloud data, and transmitting a bitstream that includes the point cloud data and signaling information. The receiving method may include receiving a bitstream that includes point cloud data and signaling information, decoding the point cloud data, and rendering the decoded point cloud data.

Abstract: The present disclosure describes systems and methods for enhancing the space utilization of storing keys in hierarchical data structures, such as B+Tree leaf blocks, by partially or fully compressing the keys. Partially compressing the keys may include storing certain keys as offsets from a start key. Fully compressing the keys may include setting a “type” value included in the header of the leaf block to a value indicating that the keys are implicitly represented by an ordinal key index, which defines the ordinal arrangement of keys in the leaf block. Partially or fully compressing keys, according to the implementations described herein, reduces the amount of space needed to store keys in a leaf block of a B+Tree.

Abstract: When operating a data processing system comprising a data encoder operable to perform a first encoding scheme that is configured for processing sets of data arranged in a first data format, for a plurality of sets of data received in a second, different data format, the bits for at least some of the received plurality of sets of data in the second data format are re-ordered to map the bits for the at least some of the received plurality of sets of data in the second data format into the first arrangement of bits associated with the first data format to thereby generate from the received plurality of sets of data in the second data format one or more sets of data in the first data format for processing using the first encoding scheme.

Abstract: Disclosed herein is a method of transmitting point cloud data. The method may include acquiring point cloud data, encoding geometry information including positions of points of the point cloud data, generating one or more levels of detail (LODs) based on the geometry information and selecting one or more neighbor points of each point to be attribute-encoded based on the one or more LODs, encoding attribute information of each point based on the selected one or more neighbor points of each point, and transmitting the encoded geometry information, the encoded attribute information, and signaling information.

Abstract: Systems, and method and computer readable media that store instructions for calculating signatures, utilizing signatures and the like.

Abstract: A device, system, and method is provided for training or prediction using a cluster-connected neural network. The cluster-connected neural network may be divided into a plurality of clusters of artificial neurons connected by weights or convolutional channels connected by convolutional filters. Within each cluster is a locally dense sub-network of intra-cluster weights or filters with a majority of pairs of neurons or channels connected by intra-cluster weights or filters that are co-activated together as an activation block during training or prediction. Outside each cluster is a globally sparse network of inter-cluster weights or filters with a minority of pairs of neurons or channels separated by a cluster border across different clusters connected by inter-cluster weights or filters. Training or predicting is performed using the cluster-connected neural network.

Abstract: A video coder can be configured to determine an intra-prediction mode for a block of video data, identify a most probable transform based on the intra-prediction mode determined for the block of video data, and code an indication of whether the most probable transform is a transform used to encode the block of video data. The most probable transform can be a non-square transform.

Abstract: Artificial image generation may include obtaining a source image, identifying quantization information from the source image, wherein identifying the quantization information includes identifying multiresolution quantization interval information from the source image, generating a restoration filtered image by restoration filtering the source image, generating a constrained restoration filtered image by constraining the restoration filtered image based on the quantization information, obtaining an unconstrained artificial image based on the constrained restoration filtered image and a generative artificial neural network obtained using a generative adversarial network, obtaining the artificial image by constraining the unconstrained artificial image based on the quantization information, and outputting the artificial image.

Abstract: A method for reconstructing a colour image acquired by a photosensitive sensor covered with a mosaic of filters of different colours making up a base pattern, obtaining the product of a demosaicing matrix with a matrix representation of a mosaic image coming from the sensor following acquisition of the colour image by the sensor, the product of the demosaicing matrix with the matrix representation of the mosaic image performing an interpolation of the colour of each pixel of the mosaic image as a function of a pixel neighbourhood of a base pattern corresponding to the base pattern of the mosaic of filters.

Abstract: Aspects of the disclosure provide a method for denoising an image. The method can include receiving an acquired image from an image acquisition system, and processing the acquired image with a nonlinear diffusion coefficient based filter having a diffusion coefficient that is calculated using gradient vector orientation information in the acquired image.

Abstract: An image compression method is provided. The image compression method includes steps of receiving, by a tile division generator, an input image; by the tile division generator, determining whether the input image corresponds to a partial updated region and whether the partial updated region coincides with one tile or a combination of tiles of a current frame and generating a determination result; and by the tile division generator, determining whether to re-perform tile division and compressing and encoding one portion of the current frame according to the determination result.

Abstract: This invention suppresses color shifts caused by quantization parameters used when encoding RAW image. An apparatus comprises a calculator which finds a color temperature of RAW image of a target frame, and calculates a gain value for each of color components in order to apply white balancing to the RAW image, a generator which generates a plurality of planes, each plane constituted by data of a single color component, from the RAW image, an adjusting unit which adjusts the white balance of each plane on the basis of the calculated gain, an encoder which encodes each plane by frequency-transforming, quantizing, and encoding each plane, and a controller which switches between encoding the plurality of planes generated by the generating unit or encoding the plurality of planes adjusted by the adjusting unit, on the basis of quantization parameters.

Abstract: The present disclosure is directed towards a device and a method for evaluating a wireless link (3) of a wireless node of a customer premises equipment (CPE) device (1, 2) during operation of the CPE device. The device and method includes determining an actual data rate (33) of the wireless link, determining an available data rate (31) of the wireless link, and calculating a head room (34) based on a required data rate (30) and the available data rate (31).

Abstract: In some embodiments, an electronic device for compressing video image data includes a housing, an image sensor, a memory device, and one or more processors. The image sensor can generate image data from light incident on the image sensor. The one or more processors can transform the image data to obtain transform coefficients, quantize the transform coefficients, encode the quantized transform coefficients, and store the quantized transform coefficients to the memory device. The one or more processors can encode the quantized transform coefficients at least by determining a range of multiple ranges in which one transform coefficient is included, determining a value within the range to which the one transform coefficient corresponds, encoding using a first algorithm the range as a range code, and encoding using a second algorithm the value as a value code.

Abstract: A recording apparatus is disclosed. The recording apparatus includes a data input portion configured to input data, a first moving image signal recording portion configured to record, based on the input data, a first moving image signal having a first image quality attribute, a condition detector configured to detect that the input data satisfies a predetermined condition during recording of the first moving image signal, and a second moving image signal recorder configured to record, based on the input data, a second moving image signal having a second image quality attribute when the condition detector detects that the input data satisfies the predetermined condition.

Abstract: Methods, apparatus, systems and articles of manufacture are disclosed to improve spatial-temporal data management. An example apparatus includes a hypervoxel data structure generator to generate a root hexatree data structure having sixteen hypernodes, an octree manager to improve a spatiotemporal data access efficiency by generating a first degree of symmetry in the root hexatree, the octree manager to assign a first portion of the hypernodes to a positive temporal subspace and to assign a second portion of the hypernodes to a negative temporal subspace, and a quadtree manager to improve the spatiotemporal data access efficiency by generating a second degree of symmetry in the root hexatree, the quadtree manager to assign respective hypernodes of the positive temporal subspace and the negative temporal subspace to respective positive and negative spatial subspaces.

Abstract: Concepts and technologies are described herein for interacting with contextual and task-focused computing environments. Tasks associated with applications are described by task data. Tasks and/or batches of tasks relevant to activities occurring at a client are identified, and a UI for presenting the tasks is generated. The UIs can include tasks and workflows corresponding to batches of tasks. Workflows can be executed, interrupted, and resumed on demand. Interrupted workflows are stored with data indicating progress, contextual information, UI information, and other information. The workflow is stored and/or shared. When execution of the workflow is resumed, the same or a different UI can be provided, based upon the device used to resume execution of the workflow. Thus, multiple devices and users can access workflows in parallel to provide collaborative task execution.

Abstract: Embodiments are generally directed to tile-based multiple resolution rendering of images. An embodiment of an apparatus includes one or more processor cores; a plurality of tiling bins, the plurality of tiling bins including a bin for each of a plurality of tiles in an image; and a memory to store data for rendering of an image in one or more of a plurality of resolutions. The apparatus is to generate, in the memory, storage for a resolution setting for each the plurality of tiling bins and storage for a final render target, each tile of the final render target being rendered based on a respective tiling bin in the plurality of tiling bins.

Abstract: Provided are a reception apparatus, a transmission apparatus, a control method, a transmission method, and a program that can control a frame image size of each frame image so that the frame image constituting a moving image is transmitted at an appropriate timing. A frame-image reception section receives a first frame image constituting the moving image. A control-data update section performs control so that the frame image size of a second frame image constituting the moving image, which is received after the first frame image, corresponds to a time required to receive the first frame image.

Abstract: A method of video decoding includes receiving a coded video bitstream including a current picture. The method further includes performing, to generate a prediction block, an optimized inverse transform process on a transform block that corresponds to a current block included in the current picture based on one of (i) a transform property of the inverse transform process and (ii) a block size of the transform block. The method further includes reconstructing the current block using the predicted block.

Abstract: An adaptive camera resolution system is described for use with matching resolution with compression. In one example, the a method includes capturing a first set of frames of a video sequence at an image sensor at a first resolution, capturing a second set of frames of the video sequence at the image sensor at a second lower resolution, wherein the first and second sets of frames are interleaved, receiving the captured frames at a video encoder, and combining the frames at the video encoder to generate an encoded video sequence of frames with the first resolution.

Abstract: The present invention includes a method for operating a data processing system to compute an approximation to a scalar product between first and second vectors in which each vector is characterized by N components. The method includes replacing the first vector by a third vector that is a pyramid integer vector characterized by N components and an integer K equal to the sum of the absolute values of the N components, and computing a scalar product of the third vector with the second vector to provide the approximation to the scalar product between the first and second vectors. Computing the scalar product of the second and third vectors can be carried out by K additions followed by one floating point multiply.

Abstract: A video recording apparatus includes an object detection unit that detects an object from an input video image; a feature amount extraction unit that extracts a feature amount of the object detected from the input video image; a correction unit that corrects the input video image; and a recording unit that, when the object is detected, records the video image corrected and the feature amount extracted in association with each other.

Abstract: The present invention is related to a method for routing image data from Camera Node to a Sink Node, where it May be processed in detail, in a Visual sensor network. In accordance with the present invention, wireless Intermediate Node are configured to route images based on first pass image encoding containing coarse image to the next wireless hop at a higher priority over the images based on the second pass encoding containing additional details of the images. The routing method for wireless Ad Hoc mobile and sensor network, periodically exchanges node status information to optimize the routing table, so that high priority image packets may reach the Sink Node in shortest possible path. The method ensures selection of high quality image packets for routing under the condition of collision and congestion to achieve optimal image quality receipt at the Sink Node.

Abstract: A data object design system and method facilitates and generates creation of data objects to contain and, thus, contain data structures within the data objects. The data objects and interaction between the data objects implement a higher level of abstraction between data objects and data entities. Particularly for complicated, sophisticated data models, the data objects support an increased usability, efficiency, and comprehension of a data modeling system. Thus, in at least one embodiment, the data object design system and method increases the technical capabilities of data modeling tools by, for example, implementing data objects to reduce complex data modeling implementations into manageable data objects and increasing the efficient utilization of data modeling tools.

Abstract: An original image containing a string of text is received. The string of text contained in the original image is determined using text recognition. A first compression level applied to the original image creates a first compressed image. A string of text contained in the first compressed image is determined using text recognition. The string of text contained in the original image is compared with the string of text contained in the first compressed image. The first compression level is incremented to reach a second compression level. The second compression level applied to the original image creates a second compressed image. A string of text contained in the second compressed is determined image using text recognition. The string of text contained in the original image is compared with the string of text contained in the second compressed image. It is determined that further image compression refinement is unnecessary.

Abstract: A method performed by an apparatus for decoding a video, includes: generating a predicted block by predicting a current block to be decoded; reconstructing frequency coefficients in a frequency conversion unit by decoding a bitstream, to generate a frequency conversion block having a size of the frequency conversion unit; inversely transforming the frequency conversion block by using a transform size identical to the size of the frequency conversion unit, to reconstruct a residual block; and adding the reconstructed residual block to the predicted block.

Abstract: A parallel processing starting unit 3 that partitions an inputted image into tiles each having a predetermined size, and distributes tiles obtained through the partitioning, and N tile encoding units 5-1 to 5-N each of that carries out a prediction difference encoding process on a tile distributed thereto by the parallel processing starting unit 3 to generate a local decoded image are disposed, and each of N tile loop filter units 7-1 to 7-N determines a filter per tile suitable for a filtering process on the local decoded image generated by the corresponding one of the tile encoding units 5-1 to 5-N, and carries out the filtering process on the local decoded image by using the filter.

Abstract: The present invention relates to an improved method of generating a passive chronological timesheet for users of computing devices having a graphical user interface.

Abstract: Techniques related to selecting restoration filter coefficients for 2-dimensional loop restoration filters for super resolution video coding are discussed. Such techniques include upscaling reconstructed video frames along only a first dimension and selecting filter coefficients for portions of the frame by an evaluation that, for each pixel of the portion, uses only pixel values that are aligned with the first dimension. Selection of filter coefficients for the second dimension may be skipped or made using only a subset of available filter coefficients.

Abstract: Methods and apparatus for processing images are provided. The image processing method includes obtaining an image using an electronic device; associating first metadata corresponding to the image with a first layer of a hierarchical data structure, wherein the hierarchical data structure includes a plurality of layers, including the first layer and a second layer; associating second metadata with the second layer, wherein the second metadata corresponds to one or more of a plurality of blocks into which the image is divided; and storing the hierarchical data structure in association with the image in a storage unit operatively coupled to the electronic device.

Abstract: An image processing device includes circuitry. The circuitry reads YUV image data for each block of a predetermined size; performs discrete wavelet transform on a Y value of a pixel included in a block to obtain a first transformed Y value; multiplies the first transformed Y value by a coefficient used for inversely transforming the first transformed Y value, to calculate a second transformed Y value; and quantizes a U value and a V value included in the block and the second transformed Y value to generate fixed length encoded data.

Abstract: Various examples are directed to methods and systems for compressing panoramic video. An image processor may receive a frame comprising a plurality of pixel values arranged according to a two-dimensional grid and divide the frame into a plurality of pixel blocks comprising a first pixel block and a second pixel block. The first pixel block may be positioned in a first edge column adjacent a first edge of the frame and the second pixel block may be positioned in a second edge column adjacent a second edge of the frame opposite the first edge. The image processor may identify a spatial continuity between the first pixel block and the second pixel block and replace the first pixel block with a reference to the second pixel block.

Abstract: A method and apparatus for storing, processing and reconstructing full resolution image out of sub band encoded images are provided. The method of rendering high resolution images based on sub-band encoded data of an image includes steps of performing downscaling of a selected image, creating a time-stamped downscaled image, extracting sub-band information associated with the selected image at one instance in time, compressing the downscaled image and the sub-band information, and decompressing and adding the sub-band information with extrapolated downscaled image to reconstruct and render high resolution image.

Abstract: An image coding method codes an image on a per-region basis. The method includes calculating offset information to be used for applying offset to the current region; determining whether or not the offset information is the same between the current region and a neighboring region of the current region; applying offset using the offset information and a result obtained in the determining, by adding an offset value to a reconstructed signal obtained by coding a pixel signal of the current region and decoding the coded pixel signal; coding the offset information; and storing, into a memory, an offset signal obtained by applying the offset, so that the offset signal is used for coding of a subsequent region.

Abstract: A method, system, and articles of high throughput arithmetic entropy coding for video coding uses a non-framewidth raster order or non-raster order to form spatial neighbor probability contexts for entropy coding.

Abstract: A parallel processing starting unit 3 that partitions an inputted image into tiles each having a predetermined size, and distributes tiles obtained through the partitioning, and N tile encoding units 5-1 to 5-N each of that carries out a prediction difference encoding process on a tile distributed thereto by the parallel processing starting unit 3 to generate a local decoded image are disposed, and each of N tile loop filter units 7-1 to 7-N determines a filter per tile suitable for a filtering process on the local decoded image generated by the corresponding one of the tile encoding units 5-1 to 5-N, and carries out the filtering process on the local decoded image by using the filter.