Abstract: One embodiment provides for a compute apparatus to perform machine learning operations, the compute apparatus comprising a decode unit to decode a single instruction into a decoded instruction, the decoded instruction to cause the compute apparatus to perform a complex machine learning compute operation.

Abstract: A method of augmented reality video communications between a remote system having a camera and a local system having a display. The remote and local systems are in communication with each other. A select portion of a remote camera video from the remote system is communicated to the local system. A local overlay is provided at the local system corresponding to a target portion of the local display field of view. Interpolation between the local overlay and the local display field of view specifies a local augmented reality portion of a local display corresponding to the local overlay. The select portion of the remote camera video is displayed onto the local augmented reality portion of the local display.

Abstract: Disclosed is a process for producing an animation of the multipage type, in which each still image of the animation has scalability properties. Preferably, each still image is to be saved on at least two levels, the first level containing few details and each following level providing additional details to the preceding level.

Abstract: An image signal processor for generating a converted image based on a raw image includes processing circuitry configured to store data corresponding to a plurality of lines of a received image in a line buffer, perform an image processing operation by filtering the data stored in the line buffer based on at least one filter, and divide the raw image into a plurality of sub-images and request the plurality of sub-images from a memory in which the raw image is stored, such that the plurality of sub-images are sequentially received by the line buffer, a width of each of the plurality of sub-images being less than a width of the line buffer, and the plurality of sub-images being parallel to each other.

Abstract: Managing data for transportation is provided via storing data at a host device, where the data is processed and encapsulated into transport units having a maximum size. The data is divided and processed into processed portions of varying size. The processed portions are stored sequentially into an output buffer to form the transport units, with a potential start field space left between each processed portion. A set of processed portions stored in the output buffer is determined based on sizes of the processed portions, the set starting after a previous actual start field where the set has a size less than the maximum size and a potential start field space after a last of the processed portions in the set is replaced by an actual start field to define the set. Sets of processed portions delimited by actual start fields may then be output.

Abstract: In a server apparatus, a first acquisition unit is configured to acquire a specific point at which the traveling route of a vehicle deviates from a guidance route. A specifying unit is configured to specify a vehicle that has passed through the acquired specific point based on traveling route information of a plurality of vehicles. A second acquisition unit is configured to acquire, from the specified vehicle, an image of the specific point captured by the vehicle. A generation unit is configured to generate information for traveling guidance relating to the specific point based on the acquired image.

Abstract: A head-mounted device can be operated with another device and/or object for which information is gathered to facilitate visual display of a representation thereof. An object can be provided with indicators that allow a head-mounted device to determine both an identity and a characteristic (e.g., position, orientation, distance, etc.) of the object. Additionally or alternatively, the head-mounted device can determine both an identity and a characteristic (e.g., position, orientation, distance, etc.) of an electronic device attached to an object for producing a virtual representation of the object. Additionally or alternatively, the head-mounted device can receive data from an electronic device attached to an object for producing a virtual representation of the object. The virtual representation of the object can resemble the physical object, even where the object itself is not independently analyzed.

Abstract: The present specification describes systems and methods for automatically generating personalized blendshapes from actor performance measurements, while preserving the semantics of a template facial animation rig. The disclosed inventions facilitate the creation of an ensemble of realistic digital double face rigs for each individual with consistent behaviour across the set with sophisticated iterative optimization techniques.

Abstract: Disclosed herein are systems and methods for detecting when geometry shaders output a constant amount of data and writing the data into an output stream buffer. In one aspect, an exemplary method comprises gathering information about a number of block executions associated with the received data, analyzing the gathered information to determine whether constant or variable amount of data is generated for at least one of: a stream output or a rasterization, and when the constant amount of data is generated for the stream output, writing the generated data directly into a stream output buffer, and when the constant amount of data is generated for the rasterization, writing the generated data into a rasterization buffer either directly or through a use of an intermediate index buffer.

Abstract: In one embodiment, a system may access a codec that encodes an appearance associated with a subject and comprise codec portions that respectively correspond to body parts of the subject. The system may generate a training codec that comprises a first subset of the codec portions (a first set of body parts) and a modified second subset of the codec portions (muted body parts). The system may decode the training codec using a machine-learning model to generate a mesh of the subject. The system may transform the mesh of the subject based on a predetermined pose. The system may update the machine-learning model based on a comparison between the transformed mesh and a target mesh of the subject having the predetermined pose. The system in the present application can train a machine-learning model to render an avatar with a pose using uncorrelated codec portions corresponding to different body parts.

Abstract: A map generation system, method and computer program product are provided to generate a shadow layer from a raster image that accurately represents the shadows of one or more buildings. In the context of a map generation system, the map generation system extracts pixel values from a raster image of one or more buildings and processes the pixel values so as to retain pixel values within a predefined range while eliminating other pixel values. The pixel values that are retained represent a shadow. The map generation system also modifies the a representation of the shadow by modifying the pixel values of respective pixels so as to have a shape corresponding to the shape of the one or more buildings. The map generation system causes presentation or storage of the building layer representing the one or more buildings and a shadow layer representing the shadow.

Abstract: Methods and systems are provided for generating, rendering, manipulating (e.g., slicing), and communicating stroke objects that form ink data. In a method of generating a stroke object, pen event data indicative of pen down, pen movement, and pen up events are sequentially received to generate point objects that collectively form a stroke object. The point objects serve as control points for interpolating curve segments. Further, a start parameter indicative of a start point within a starting curve segment of the curve segments at which display of the stroke object starts, and an end parameter indicative of an end point within an ending curve segment of at which display of the stroke object ends, are generated for the stroke object. When rendering the stroke object, a system limits display of the stroke object to a range bound by the start and end parameters, as opposed to displaying the entire stroke object.

Abstract: An operating method of a display driving integrated circuit (DDIC) includes correcting first image data using correction data in a first operating mode. The first image data is received from a processor via an interface, and the correction data is stored in a first memory included in the DDIC. The method further includes storing second image data received from the processor in the first memory in response to a mode switching signal controlling the DDIC to switch to a second operating mode, and displaying the second image data on a display panel in the second operating mode.

Abstract: In a method of operating a display device supporting a variable frame mode, frame data are received during a constant active period of a frame period including the active period and a variable blank period, the received frame data are written to a frame memory in the active period, the received frame data are outputted to a data driver in the active period to display an image based on the received frame data, a time of the variable blank period is counted, and, when the time of the variable blank period reaches a predetermined threshold blank time, the frame data stored in the frame memory are outputted to the data driver in the variable blank period to display an image based on the frame data stored in the frame memory.

Abstract: A method, system and computer program product are configured to analyze an image of a vehicle to determine a characteristic of the vehicle, such as may be represented by or otherwise at least partially defined by the shadow cast by the vehicle. In the context of a method, information is received identifying a vehicle from a raster image and the pixel values of the raster image are evaluated to identify pixels having pixel values representative of a shadow associated with the vehicle. The method also modifies a representation of the shadow by modifying the pixel values of the pixels based upon a shape of the vehicle such that the representation of the shadow, as modified, has a shape corresponding to the shape of the vehicle. The method additionally determines a characteristic of the vehicle based upon the representation of the shadow, as modified, that is associated with the vehicle.

Abstract: The present invention discloses a method for real-time rendering of giga-pixel images. Image data are subject to offline pre-processing, and then are subject to data decoding and redirection through a decoding module. A corresponding scheduling strategy is determined according to different image inputs, and rendering is executed by a renderer. The present invention may realize the real-time rendering of a giga-pixel panoramic view in a conventional display device, to greatly reduce the resource allocated for rendering of giga-pixel images. The present invention may render an image originally requiring a 40G or more video memory capacity on a common video card with a 1G-4G video memory capacity.

Abstract: Embodiments described herein are generally directed to conservative rasterization pipeline configurations that allow EarlyZ to be enabled for conservative rasterization.

Abstract: Example systems and methods for virtual visualization of a three-dimensional (3D) model of an object in a two-dimensional (2D) environment. The method may include capturing the 2D environment and adding scale and perspective to the 2D environment. Further, a user may select intersection points on a ground plane of the 2D environment to form walls, thereby converting the 2D environment into a 3D space. The user may further add 3D models of objects on the wall plane such that the objects may remain flush with the wall plane.

Abstract: Methods, systems, and devices for deep learning based head motion prediction for extended reality are described. The head pose prediction may involve training one or more layers of a machine learning network based on application data and an estimated head motion range associated with the extended reality system. The network may receive one or more bias corrected inertial measurement unit (IMU) measurements based on a sensor. The network may model a relative head pose of the user as a polynomial of time over a prediction interval based on the bias corrected IMU measurements and the trained one or more layers of the machine learning network. The network may determine a future relative head pose of the user based on the polynomial (e.g., which may be used for virtual object generation, display, etc. within an extended reality system).

Abstract: A photo filter (e.g., artistic) light field effect system comprises an eyewear device that includes a frame, a temple connected to a lateral side of the frame, and a depth-capturing camera. Execution of programming by a processor configures the photo filter light field effect system to apply a photo filter selection to: (i) a left raw image or a left processed image to create a left photo filter image, and (ii) a right raw image or a right processed image to create a right photo filter image. The photo filter light field effect system generates, a photo filter light field effect image with an appearance of a spatial rotation or movement, by blending together the left photo filter image and the right photo filter image based on a left image disparity map and a right image disparity map.