Abstract: Various implementations disclosed herein include devices, systems, and methods that dynamically apply a 3D effect to a 2D asset. For example, a process may obtain an image depicting two-dimensional (2D) content. The process may further determine to apply a three-dimensional (3D) effect to the image via a head mounted device (HMD). The process may further, in accordance with determining to apply the 3D effect to the image, present a view of a 3D environment including the image. The image may be positioned at a location within the 3D environment and the view may depict the image using the 3D effect.

Abstract: A mobile communication terminal device may include one or more image sensors, configured to generate image sensor data representing an environment of the mobile communication terminal device; one or more processors, configured to receive the image sensor data from the one or more image sensors; implement at least one artificial neural network to receive the image sensor data as an artificial neural network input and output an artificial neural network output representing a detected environment parameter of the environment of the mobile communication terminal; determine a navigation instruction based on the artificial neural network output; and send a signal representing the navigation instruction to a robot terminal via a communication interface.

Abstract: Apparatus and method for enhancing graphics rendering photorealism. For example, one embodiment of a graphics processor comprises: a graphics processing pipeline comprising a plurality of graphics processing stages to render a graphics image; a local storage to store intermediate rendering data to generate the graphics image; and machine-learning hardware logic to perform a refinement operation on the graphics image using at least a portion of the intermediate rendering data to generate a translated image.

Abstract: Described herein are techniques for learning neural reflectance shaders from images. A set of one or more machine learning models can be trained to optimize an illumination latent code and a set of reflectance latent codes for an object within a set of input images. A shader can then be generated based on a machine learning model of the one or more machine learning models. The shader is configured to sample the illumination latent code and the set of reflectance latent codes for the object. A 3D representation of the object can be rendered using the generated shader.

Abstract: Systems, apparatuses and methods may provide for technology that generates, by a first neural network, an initial set of model weights based on input data and iteratively generates, by a second neural network, an updated set of model weights based on residual data associated with the initial set of model weights and the input data. Additionally, the technology may output a geometric model of the input data based on the updated set of model weights. In one example, the first neural network and the second neural network reduce the dependence of the geometric model on the number of data points in the input data.

Abstract: Described herein are techniques for learning neural reflectance shaders from images. A set of one or more machine learning models can be trained to optimize an illumination latent code and a set of reflectance latent codes for an object within a set of input images. A shader can then be generated based on a machine learning model of the one or more machine learning models. The shader is configured to sample the illumination latent code and the set of reflectance latent codes for the object. A 3D representation of the object can be rendered using the generated shader.

Abstract: Systems, apparatuses and methods may provide for technology that estimates poses of a plurality of input images, reconstructs a proxy three-dimensional (3D) geometry based on the estimated poses and the plurality of input images, detects a user selection of a virtual viewpoint, encodes, via a first neural network, the plurality of input images with feature maps, warps the feature maps of the encoded plurality of input images based on the virtual viewpoint and the proxy 3D geometry, and blends, via a second neural network, the warped feature maps into a single image, wherein the first neural network is deep convolutional network and the second neural network is a recurrent convolutional network.

Abstract: Some embodiments are directed to a neural network training device for training a neural network. At least one layer of the neural network layers is a projection layer. The projection layer projects a layer input vector (x) of the projection layer to a layer output vector (y). The output vector (y) sums to the summing parameter (k).

Abstract: Systems, apparatuses and methods may provide for technology that generates, by a first neural network, an initial set of model weights based on input data and iteratively generates, by a second neural network, an updated set of model weights based on residual data associated with the initial set of model weights and the input data. Additionally, the technology may output a geometric model of the input data based on the updated set of model weights. In one example, the first neural network and the second neural network reduce the dependence of the geometric model on the number of data points in the input data.

Abstract: Described herein are techniques for learning neural reflectance shaders from images. A set of one or more machine learning models can be trained to optimize an illumination latent code and a set of reflectance latent codes for an object within a set of input images. A shader can then be generated based on a machine learning model of the one or more machine learning models. The shader is configured to sample the illumination latent code and the set of reflectance latent codes for the object. A 3D representation of the object can be rendered using the generated shader.

Abstract: Methods, apparatus, systems and articles of manufacture disclosed herein perform dense prediction of an input image using transformers at an encoder stage and at a reassembly stage of an image processing system. A disclosed apparatus includes an encoder with an embedder to convert an input image to a plurality of tokens representing features extracted from the input image. The tokens are embedded with a learnable position embedding. The encoder also includes one or more transformers configured in a sequence of stages to relate the tokens to each other. The apparatus further includes a decoder that includes one or more of reassemblers to assemble the tokens into feature representations, one or more of fusion blocks to combine the feature representations to generate a final feature representation, and an output head to generate a dense prediction based on the final feature representation and based on an output task.

Abstract: Systems and methods for operating a deep equilibrium (DEQ) model in a neural network are disclosed. DEQs solve for a fixed point of a single nonlinear layer, which enables decoupling the internal structure of the layer from how the fixed point is actually computed. This disclosure discloses that such decoupling can be exploited while substantially enhancing this fixed point computation using a custom neural solver.

Abstract: A computer-implemented method for a classification and training a neural network includes receiving input at the neural network, wherein the input includes a plurality of resolution inputs of varying resolutions, outputting a plurality of feature tensors for each corresponding resolution of the plurality of resolution inputs, fusing the plurality of feature tensors utilizing upsampling or down sampling for the vary resolutions, utilizing an equilibrium solver to identify one or more prediction vectors from the plurality of feature tensors, and outputting a loss in response to the one or more prediction vectors.

Abstract: Regularized training of a Deep Equilibrium Model (DEQ) is provided. A regularization term is computed using a predefined quantity of random samples and the Jacobian matrix of the DEQ, the regularization term penalizing the spectral radius of the Jacobian matrix. The regularization term is included in an original loss function of the DEQ to form a regularized loss function. A gradient of the regularized loss function is computed with respect to model parameters of the DEQ. The gradient is used to update the model parameters.

Abstract: Methods, apparatus, systems, and articles of manufacture are disclosed for metric depth estimation using a monocular visual-inertial system. An example apparatus for metric depth estimation includes at least one memory, instructions in the apparatus, and processor circuitry to execute the instructions to access a globally-aligned depth prediction, the globally-aligned depth prediction generated based on a monocular depth estimator, access a dense scale map scaffolding, the dense scale map scaffolding generated based on visual-inertial odometry, regress a dense scale residual map determined using the globally-aligned depth prediction and the dense scale map scaffolding, and apply the dense scale residual map to the globally-aligned depth prediction.

Abstract: Systems, apparatuses and methods may provide for technology that generates, by a first neural network, an initial set of model weights based on input data and iteratively generates, by a second neural network, an updated set of model weights based on residual data associated with the initial set of model weights and the input data. Additionally, the technology may output a geometric model of the input data based on the updated set of model weights. In one example, the first neural network and the second neural network reduce the dependence of the geometric model on the number of data points in the input data.

Abstract: Systems, apparatuses and methods may provide for technology that generates, by a first neural network, an initial set of model weights based on input data and iteratively generates, by a second neural network, an updated set of model weights based on residual data associated with the initial set of model weights and the input data. Additionally, the technology may output a geometric model of the input data based on the updated set of model weights. In one example, the first neural network and the second neural network reduce the dependence of the geometric model on the number of data points in the input data.

Abstract: Systems, apparatuses and methods may provide for technology that selects elements of a multi-scale kernel according to resolutions in an adaptive grid, conducts convolutions on the adaptive grid with the selected elements of the multi-scale kernel, and generates a signed distance field based on the convolutions.

Abstract: Apparatus and method for enhancing graphics rendering photorealism. For example, one embodiment of a graphics processor comprises: a graphics processing pipeline comprising a plurality of graphics processing stages to render a graphics image; a local storage to store intermediate rendering data to generate the graphics image; and machine-learning hardware logic to perform a refinement operation on the graphics image using at least a portion of the intermediate rendering data to generate a translated image.

Abstract: Methods, apparatus, systems and articles of manufacture disclosed herein perform dense prediction of an input image using transformers at an encoder stage and at a reassembly stage of an image processing system. A disclosed apparatus includes an encoder with an embedder to convert an input image to a plurality of tokens representing features extracted from the input image. The tokens are embedded with a learnable position embedding. The encoder also includes one or more transformers configured in a sequence of stages to relate the tokens to each other. The apparatus further includes a decoder that includes one or more of reassemblers to assemble the tokens into feature representations, one or more of fusion blocks to combine the feature representations to generate a final feature representation, and an output head to generate a dense prediction based on the final feature representation and based on an output task.