According to embodiments, a scene modelling system may (e.g., initially) obtain and (e.g., subsequently) update a model of a scene based on data describing the scene. The data describing the scene may be received from any of sensors and objects, for example, located in the scene. The scene may comprise a set of connected and unconnected objects. An object may be associated with its own part of the model that may have been built, for example in an initialization phase. A connected object may transmit its (e.g., part of) model to the scene modelling system (e.g., on demand or upon detection of any change). An unconnected object (e.g., and its status) may be recognized in the scene from an image of the object, for example, captured in the scene.

Embodiments of a learning-based excavation planning method are disclosed for excavating rigid objects in clutter, which is challenging due to high variance of geometric and physical properties of objects, and large resistive force during the excavation. A convolutional neural network is utilized to predict a probability of excavation success. Embodiments of a sampling-based optimization method are disclosed for planning high-quality excavation trajectories by leveraging the learned prediction model. To reduce simulation-to-real gap for excavation learning, voxel-based representations of an excavation scene are used. Excavation experiments were performed in both simulation and real world to evaluate the learning-based excavation planners. Experimental results show that embodiments of the disclosed method may plan high-quality excavations for rigid objects in clutter and outperform baseline methods by large margins.

Even under a situation where an object to be presented has movement, the object can be presented as display information in a more favorable mode. An information processing apparatus includes an acquisition unit (111) that acquires information regarding movement of an object, and a control unit (111) that projects the object in a display region at a projection timing set according to a first period, and corrects display information according to a result of the projection in accordance with a plurality of display timings each set for each second period shorter than the first period, and the control unit controls correction of second display information to maintain continuity according to the movement of the object between first display information displayed according to a first projection result of the object in accordance with a first display timing, and the second display information displayed according to a second projection result of the object in accordance with a second display timing immediately after the first display timing.

A generation device includes processing circuitry configured to receive a plurality of inputs of road map data including longitude/latitude data on lane information indicating a center line of a lane, longitude/latitude data on a road shoulder line, and longitude/latitude data on a lane marker, and refer to the road map data and generate a first polygon indicating a region of a lane based on a plurality of intersections on the lane marker or the road shoulder line crossed by a vertical line from the lane information.

Methods and systems for providing landing area information on an active avionic display on a display device in a cockpit of an aircraft. The system includes an avionic display module configured to render an avionic display on a display device; and a landing area guidance module configured to: construct a graphical insert that is smaller than the avionic display, the graphical insert depicting the landing area environment as a two-dimensional area that includes a landing location rendered therein in a first visualization scheme; overlay the graphical insert on a small portion of the avionic display; indicate a target exit on the landing location; determine whether a runway overrun awareness alerting system (ROAAS) alert has been received; determine whether a surface indication alerts (SurfIA) alert has been received; responsive to the ROAAS alert and the SurfIA alert, alter the rendering within the graphical insert.

In various examples, a generative model is used to synthesize datasets for use in training a downstream machine learning model to perform an associated task. The synthesized datasets may be generated by sampling a scene graph from a scene grammar—such as a probabilistic grammar—and applying the scene graph to the generative model to compute updated scene graphs more representative of object attribute distributions of real-world datasets. The downstream machine learning model may be validated against a real-world validation dataset, and the performance of the model on the real-world validation dataset may be used as an additional factor in further training or fine-tuning the generative model for generating the synthesized datasets specific to the task of the downstream machine learning model.

An application sends primitives to a graphics processing system so that an image of a 3D scene can be rendered. The primitives are placed into primitive blocks for storage and retrieval from a parameter memory. Rather than simply placing the first primitives into a primitive block until the primitive block is full and then placing further primitives into the next primitive block, multiple primitive blocks can be “open” such that a primitive block allocation module can allocate primitives to one of the open primitive blocks to thereby sort the primitives into primitive blocks according to their spatial positions. By grouping primitives together into primitive blocks in accordance with their spatial positions, the performance of a rasterization module can be improved. For example, in a tile-based rendering system this may mean that fewer primitive blocks need to be fetched by a hidden surface removal module in order to process a tile.

A method is provided that includes generating a four-dimensional (4D) model of an environment based on three-dimensional (3D) coordinates of the environment captured at a first point in time. The method further includes updating the 4D model based at least in part to an update to at least a subset of the 3D coordinates of the environment captured at a second point in time. The method further includes enriching the 4D model by adding supplemental information to the model.

Systems and methods for generating a video of a user-defined virtual reality scene are disclosed. Exemplary implementations may: obtain a scene definition; obtain camera information for multiple virtual cameras to be used in generating a two-dimensional presentation of the virtual reality scene; execute a simulation of the virtual reality scene from the scene definition for at least a portion of the scene duration; obtain camera timing instructions specifying which of the virtual cameras should be used to generate the two-dimensional presentation of the virtual reality scene as a function of progress through the scene duration; generate the two-dimensional presentation of the virtual reality scene in accordance with the camera timing instructions and the camera information.

This technology relates to rendering content from discrete applications. In this regard, one or more computing devices may receive a global scene graph containing resources provided by two or more discrete processes, wherein the global scene graph is instantiated by a first process of the two or more discrete processes. The one or more computing devices may render and output for display, the global scene graph in accordance with the resources contained there.