Abstract: Systems and methods are provided for providing an audio-guided in-door virtual reality (VR) tour. An exemplary system may include a communication interface configured to receive input from a user and to output media contents, a memory storing computer-readable instructions, and at least one processor coupled to the communication interface and the memory. The computer-readable instructions, when executed by the processor, may cause the at least one processor to perform operations. The operations may include displaying a view of a 3D VR environment and playing an audio guide associated with the view. The operations may also include detecting, during the playing of the audio guide, a target operation input by the user to alter the view. In response to the detection of the target operation, the operations may include adjusting, based on the detected target operation, the view with respect to a fixed point position within the 3D VR environment.

Abstract: In one embodiment, a method includes receiving instructions to render a snapshot of a scene for a video, where the snapshot is to be displayed using a sequence of N frames, computing a mipmap-level determining factor for a texture appearing in the scene based on a scale of the texture on a pixel grid, selecting a mipmap level of the texture for each of the N frames based on the mipmap-level determining factor, where the mipmap levels selected for the N frames are non-uniform and temporally approximate the mipmap-level determining factor, rendering each of the N frames by sampling the mipmap level of the texture selected for that frame, and displaying the rendered N frames sequentially to represent the snapshot of the scene.

Abstract: During tracing of a primary ray in a 3-D space (e.g., a 3-D scene in graphics rendering), a ray is found to intersect a primitive (e.g., a triangle) located in the 3-D space. Secondary ray(s) may be generated for a variety of purposes. For example, occlusion rays may be generated to test occlusion of a point of intersection between the primary ray and primitive is illuminated by any of the light(s). An origin for each secondary ray can be modified from the intersection point based on characteristics of the primitive intersected. For example, an offset from the intersection point can be calculated using barycentric coordinates of the intersection point and interpolation of one or more parameters associated with vertices defining the primitive. These parameters may include a size of the primitive and differences between a geometric normal for the primitive and a respective additional vector supplied with each vertex.

Abstract: Systems, methods, and/or computer-readable media described herein provide technical solutions to the highly technical problems of machine generation of dental restorations. In particular, these systems, methods and/or computer readable media may provide technical solutions to aid in the creation of dental restorations that more closely resemble a natural tooth (including its internal optical structure). These systems, methods and/or computer readable media may help in virtually rendering a tooth, including its internal optical structure, and apply these renderings (e.g., digital models) to the fabrication of the dental restoration.

Abstract: A fully-connected neural network may be configured for execution by a processor as a fully-fused neural network by limiting slow global memory accesses to reading and writing inputs to and outputs from the fully-connected neural network. The computational cost of fully-connected neural networks scale quadratically with its width, whereas its memory traffic scales linearly. Modern graphics processing units typically have much greater computational throughput compared with memory bandwidth, so that for narrow, fully-connected neural networks, the linear memory traffic is the bottleneck. The key to improving performance of the fully-connected neural network is to minimize traffic to slow “global” memory (off-chip memory and high-level caches) and to fully utilize fast on-chip memory (low-level caches, “shared” memory, and registers), which is achieved by the fully-fused approach.

Abstract: An automated cage-to-cage fitting technique is used to fit arbitrary three-dimensional (3D) geometry to arbitrary target 3D geometry for 3D avatars in a 3D environment, such as an online game. Each graphical representation of items of clothing is provided with an inner cage and with an outer cage. The inner cage of a clothing item (a current layer) is mapped to the outer cage of another clothing item (a previous/underlying layer) that has already been fitted on the 3D avatar. This mapping enables the current layer to conform to the shape of the previous layer. The technique allows any body geometry to be fitted with any clothing geometry, including enabling layers of clothing to be fitted over underlying layer(s) of clothing, thereby providing customization.

Abstract: A method and intersection testing module in a ray tracing system for determining whether a ray intersects a 3D axis-aligned box that represents a volume defined by a front-facing plane and a back-facing plane for each dimension. Scaled inverse ray components are determined and a scaled minimum culling distance is determined using a result of multiplying an unscaled minimum culling distance for the ray by a predetermined magnitude. Scaled intersection distances to the planes defining the box are determined using scaled inverse ray components. The largest of the determined scaled intersection distances to a front-facing plane of the box is identified. The smallest of the determined scaled intersection distances to a back-facing plane of the box is identified. It is determined that the ray intersects the box if all of three determinations are satisfied, and it is determined that the ray misses the box if one or more of the three determinations are not satisfied.

Abstract: A computing device is provided, comprising a processor configured to execute a physics engine. The physics engine is configured to, during narrowphase collision detection of a collision detection phase, identify a set of convex polyhedron pairs, each including a first convex polyhedron from a first rigid body and a second convex polyhedron from a second rigid body. The physics engine is further configured to, for each convex polyhedron pair, determine a separating plane. The physics engine is further configured to perform neighbor welding on pair combinations of the convex polyhedron pairs during the narrowphase collision detection to thereby modify the separating planes of at least a subset of the convex polyhedron pairs. The physics engine is further configured to determine collision manifolds for the convex polyhedron pairs, including for the subset of convex polyhedron pairs having the modified separating planes.

Abstract: Systems and methods for planning delivery of an object via a catheter, such as transseptal delivery of a prosthetic mitral valve to a patient's heart are disclosed.

Abstract: Technologies are provided herein for modeling and tracking physical objects, such as human hands, within a field of view of a depth sensor. A sphere-mesh model of the physical object can be created and used to track the physical object in real-time. The sphere-mesh model comprises an explicit skeletal mesh and an implicit convolution surface generated based on the skeletal mesh. The skeletal mesh parameterizes the convolution surface and distances between points in data frames received from the depth sensor and the sphere-mesh model can be efficiently determined using the skeletal mesh. The sphere-mesh model can be automatically calibrated by dynamically adjusting positions and associated radii of vertices in the skeletal mesh to fit the convolution surface to a particular physical object.

Abstract: Techniques of compressing level of detail (LOD) data involve sharing a texture image LOD among different mesh LODs for single-rate encoding. That is, a first texture image LOD corresponding to a first mesh LOD may be derived by refining a second texture image LOD corresponding to a second mesh LOD. This sharing is possible when texture atlases of LOD meshes are compatible.

Abstract: A computer-implemented method for generating a virtual depiction of an orthodontic treatment of a patient is disclosed herein. The computer-implemented method may involve gathering a three-dimensional (3D) model modeling the patient's dentition at a specific treatment stage of an orthodontic treatment plan. An image of the patient's face and dentition may be gathered. A first set of reference points modeled on the 3D model of the patient's dentition and a second set of reference points represented on the dentition of the image of the patient may be received. The image of the patient's dentition may be projected into a 3D space to create a projected 3D model of the image of the patient's dentition. Based on a comparison of the first reference points and projections of the second set of reference points, a plurality of modified images of the patient may be constructed to depict progressive stages of a treatment plan.

Abstract: The present disclosure provides systems and methods for generating photorealistic image simulation data with geometry-aware composition for testing autonomous vehicles. In particular, aspects of the present disclosure can involve the intake of data on an environment and output of augmented data on the environment with the photorealistic addition of an object. As one example, data on the driving experiences of a self-driving vehicle can be augmented to add another vehicle into the collected environment data. The augmented data may then be used to test safety features of software for a self-driving vehicle.

Abstract: A cross reality system enables any of multiple devices to efficiently and accurately access previously persisted maps, even maps of very large environments, and render virtual content specified in relation to those maps. The cross reality system may quickly process a batch of images acquired with a portable device to determine whether there is sufficient consistency across the batch in the computed localization. Processing on at least one image from the batch may determine a rough localization of the device to the map. This rough localization result may be used in a refined localization process for the image for which it was generated. The rough localization result may also be selectively propagated to a refined localization process for other images in the batch, enabling rough localization processing to be skipped for the other images.

Abstract: A method includes the steps of receiving training data comprising images of an object and associated camera poses from which the images are captured, training, based on the training data, a machine-learning model to take as input a given viewpoint and synthesize an image of a virtual representation of the object viewed from the given viewpoint, generating, for each of predetermined viewpoints surrounding the virtual representation of the object, a view-dependent image of the object as viewed from that viewpoint using the trained machine-learning model, receiving, from a client device, a desired viewpoint from which to view the virtual representation of the object, selecting one or more of the predetermined viewpoints based on the desired viewpoint, and sending, to the client device, the view-dependent images associated with the selected one or more viewpoints for rendering an output image of the virtual representation of the object viewed from the desired viewpoint.

Abstract: A cross reality system receives tracking information in a tracking map and first location metadata associated with at least a portion of the tracking map. A sub-portion of a canonical map is determined based at least in part on a correspondence between the first location metadata associated with the at least the portion of the tracking map and second location metadata associated with the sub-portion of the canonical map. The sub-portion of the canonical map may be merged with the at least the portion of the tracking map into a merged map. The cross reality system may further generate the tracking map by using at least the pose information from one or more images and localize the tracking map to the canonical may at least by using a persistent coordinate frame in the canonical map and the location metadata associated with the location represented in the tracking map.

Abstract: In some implementations, a method of improved pathfinding is performed at a virtual agent operating system including non-transitory memory and one or more processors coupled with the non-transitory memory. The method includes: determining an initial path for a virtual agent to a target destination based at least in part on a navigation mesh of an XR environment; actuating locomotive elements of the virtual agent in order to move the virtual agent according to the initial path; while moving according to the initial path, detecting a node of a navigation graph; in response to detecting the node of the navigation graph: obtaining navigation information from the node of the navigation graph; and determining an updated path from the node to the target destination based at least in part on the navigation mesh and the navigation information; and actuating the locomotive elements of the virtual agent according to the updated path.

Abstract: A method and system for culling a patch of surface data from one or more tiles in a tile based computer graphics system. A rendering space is divided into a plurality of tiles and a patch of surface data read. Then, at least a portion of the patch is analysed to determine data representing a bounding depth value evaluated over at least one tile. This may comprise tessellating the patch of surface data to derive a plurality of tessellated primitives and analysing at least some of the tessellated primitives. For each tile within which the patch is located, the data representing the bounding depth value is then used to determine whether the patch is hidden in the tile, and at least a portion of the patch is rendered, if the patch is determined not to be hidden in at least one tile.

Abstract: In one embodiment, a method includes determining correspondence data between a sequence of images based on identified features in the sequence of images and predicted pose based on motion data, and determining current state information based on the correspondence data and the motion data. The current state information comprises at least a current pose of the wearable device relative to the environment capture by the one or more cameras. Furthermore, the method comprises receiving map points in a three-dimensional map and their associated descriptors based on the identified features in the sequence of images and identifying one or more of the map points in the sequence of images based on the associated descriptors associated with the map points. The current state information is further determined based on the identified one or more of the map points.

Abstract: Interactive systems using adaptive control software and hardware from known and later developed eyepieces to later developed head-wear to lenses, including implantable, temporarily insertable and contact and related film based types of lenses including thin film transparent elements for housing cameras lenses and projector and functional equivalent processing tools. Simple controls, real-time updates and instant feedback allow implicit optimization of a universal model while managing complexity.