Abstract: A neural network-based rendering technique increases temporal stability and image fidelity of low sample count path tracing by optimizing a distribution of samples for rendering each image in a sequence. A sample predictor neural network learns spatio-temporal sampling strategies such as placing more samples in dis-occluded regions and tracking specular highlights. Temporal feedback enables a denoiser neural network to boost the effective input sample count and increases temporal stability. The initial uniform sampling step typically present in adaptive sampling algorithms is not needed. The sample predictor and denoiser operate at interactive rates to achieve significantly improved image quality and temporal stability compared with conventional adaptive sampling techniques.

Abstract: A method, computer readable medium, and system are disclosed for temporally stable data reconstruction. A sequence of input data including artifacts is received. A first input data frame is processed using layers of a neural network model to produce external state including a reconstructed first data frame that approximates the first input data frame without artifacts. Hidden state generated during processing of the first input data is not provided as an input to the layer to process second input data. The external state is warped, using difference data corresponding to changes between input data frames, to produce warped external state more closely aligned with the second input data frame. The second input data frame is processed, based on the warped external state, using the layers of the neural network model to produce a reconstructed second data frame that approximates the second data frame without artifacts.

Abstract: A neural network-based rendering technique increases temporal stability and image fidelity of low sample count path tracing by optimizing a distribution of samples for rendering each image in a sequence. A sample predictor neural network learns spatio-temporal sampling strategies such as placing more samples in dis-occluded regions and tracking specular highlights. Temporal feedback enables a denoiser neural network to boost the effective input sample count and increases temporal stability. The initial uniform sampling step typically present in adaptive sampling algorithms is not needed. The sample predictor and denoiser operate at interactive rates to achieve significantly improved image quality and temporal stability compared with conventional adaptive sampling techniques.

Abstract: A neural network-based rendering technique increases temporal stability and image fidelity of low sample count path tracing by optimizing a distribution of samples for rendering each image in a sequence. A sample predictor neural network learns spatio-temporal sampling strategies such as placing more samples in dis-occluded regions and tracking specular highlights. Temporal feedback enables a denoiser neural network to boost the effective input sample count and increases temporal stability. The initial uniform sampling step typically present in adaptive sampling algorithms is not needed. The sample predictor and denoiser operate at interactive rates to achieve significantly improved image quality and temporal stability compared with conventional adaptive sampling techniques.

Abstract: In various examples, systems and methods are disclosed herein for dynamically updating a neural network having a plurality of kernels. The system may identify a first subset of kernels from the plurality of kernels in the neural network. The system may then determine the characteristics of each respective kernel in the first subset. The system may then compare the characteristics of the respective kernels in the first subject to a dynamic rule set. In response to the system comparing the characteristics of the respective kernels in the first subset to the dynamic rule set, the system identifies a second subset of the first subset based on the comparing, automatically generates instructions to combine the second subset of kernels, and updates the neural network based on the one or more instructions. The neural network may have a simplified compute graph based on the above dynamic updating systems and methods.

Abstract: A neural network-based rendering technique increases temporal stability and image fidelity of low sample count path tracing by optimizing a distribution of samples for rendering each image in a sequence. A sample predictor neural network learns spatio-temporal sampling strategies such as placing more samples in dis-occluded regions and tracking specular highlights. Temporal feedback enables a denoiser neural network to boost the effective input sample count and increases temporal stability. The initial uniform sampling step typically present in adaptive sampling algorithms is not needed. The sample predictor and denoiser operate at interactive rates to achieve significantly improved image quality and temporal stability compared with conventional adaptive sampling techniques.

Abstract: A neural network-based rendering technique increases temporal stability and image fidelity of low sample count path tracing by optimizing a distribution of samples for rendering each image in a sequence. A sample predictor neural network learns spatio-temporal sampling strategies such as placing more samples in dis-occluded regions and tracking specular highlights. Temporal feedback enables a denoiser neural network to boost the effective input sample count and increases temporal stability. The initial uniform sampling step typically present in adaptive sampling algorithms is not needed. The sample predictor and denoiser operate at interactive rates to achieve significantly improved image quality and temporal stability compared with conventional adaptive sampling techniques.

Abstract: A neural network-based rendering technique increases temporal stability and image fidelity of low sample count path tracing by optimizing a distribution of samples for rendering each image in a sequence. A sample predictor neural network learns spatio-temporal sampling strategies such as placing more samples in dis-occluded regions and tracking specular highlights. Temporal feedback enables a denoiser neural network to boost the effective input sample count and increases temporal stability. The initial uniform sampling step typically present in adaptive sampling algorithms is not needed. The sample predictor and denoiser operate at interactive rates to achieve significantly improved image quality and temporal stability compared with conventional adaptive sampling techniques.

Abstract: A method, computer readable medium, and system are disclosed for temporally stable data reconstruction. A sequence of input data including artifacts is received. A first input data frame is processed using layers of a neural network model to produce external state including a reconstructed first data frame that approximates the first input data frame without artifacts. Hidden state generated during processing of the first input data is not provided as an input to the layer to process second input data. The external state is warped, using difference data corresponding to changes between input data frames, to produce warped external state more closely aligned with the second input data frame. The second input data frame is processed, based on the warped external state, using the layers of the neural network model to produce a reconstructed second data frame that approximates the second data frame without artifacts.

Abstract: A three dimensional (3D) modeling system for generating a 3D representation of a modeled object on a display device of a computer system. The modeled object is represented by an initial definition of an object and a set of modifiers. Each modifier modifies some portion of the definition of an object that may result in a change in appearance of the object when rendered. The modifiers are ordered so that the first modifier modifies some portion of the initial definition of the object and produces a modified definition. The next modifier modifies the results of the previous modifier. The results of the last modifier are then used in rendering processes to generate the 3D representation. Each modifier is associated with a three dimensional representation so that the user can more easily visualize the effect of the modifier.

Abstract: A three dimensional (3D) modeling system for generating a 3D representation of a modeled object on a display device of a computer system. The modeled object is represented by an initial definition of an object and a set of modifiers. Each modifier modifies some portion of the definition of an object that may result in a change in appearance of the object when rendered. The modifiers are ordered so that the first modifier modifies some portion of the initial definition of the object and produces a modified definition. The next modifier modifies the results of the previous modifier. The results of the last modifier are then used in rendering processes to generate the 3D representation. Each modifier is associated with a three dimensional representation so that the user can more easily visualize the effect of the modifier.

Abstract: An interactive rendering system which can minimize computational demand while allowing a designer to manipulate one or more selected objects in a scene is disclosed. A scene is rendered to a scene buffer. One or more objects are selected and rendered to an object buffer. The scene is re-rendered to the scene buffer without the selected objects. As the selected objects move or change, they are re-rendered only in the object buffer and a display is generated by merging the objects buffer and the scene buffer. Because there is no need to render the background scene, most of the computational power can be dedicated to the selected objects. The perspective and depth relationship between the selected objects and the scene are maintained.

Abstract: A three dimensional (3D) modeling system for generating a 3D representation of a modeled object on a display device of a computer system. The modeled object is represented by an initial definition of an object and a set of modifiers. Each modifier modifies some portion of the definition of an object that may result in a change in appearance of the object when rendered. The modifiers are ordered so that the first modifier modifies some portion of the initial definition of the object and produces a modified definition. The next modifier modifies the results of the previous modifier. The results of the last modifier are then used in rendering processes to generate the 3D representation.

Abstract: A system to permit a designer to select the minimum scene refresh rate or animation redraw rate acceptable by the designer. The system also incorporates a selection of presentation methods for the three-dimensional objects in a scene, arranged in decreasing fidelity, and consequently in computational complexity, as fall back positions for the general rendering technique. Using the designer's selected refresh rate as a target, the system attempts to use the highest quality rendering technique selected by the designer. If the computational complexity of the scene causes the refresh rate of the scene to fall below the minimum acceptable level selected by the designer, the system selects the next lower rendering option selected by the designer, thus degrading the presentation quality of the scene and simultaneously reducing the computational complexity of the task.

Abstract: A three dimensional (3D) modeling system for generating a 3D representation of a modeled object on a display device of a computer system. The modeled object is represented by an initial definition of an object and a set of modifiers. Each modifier modifies some portion of the definition of an object that may result in a change in appearance of the object when rendered. The modifiers are ordered so that the first modifier modifies some portion of the initial definition of the object and produces a modified definition. The next modifier modifies the results of the previous modifier. The results of the last modifier are then used in rendering processes to generate the 3D representation. Each modifier is associated with a three dimensional representation so that the user can more easily visualize the effect of the modifier.

Abstract: A three dimensional (3D) modeling system for generating a 3D representation of a modeled object on a display device of a computer system. The modeled object is represented by an initial definition of an object and a set of modifiers. Each modifier modifies some portion of the definition of an object that may result in a change in appearance of the object when rendered. The modifiers are ordered so that the first modifier modifies some portion of the initial definition of the object and produces a modified definition. The next modifier modifies the results of the previous modifier. The results of the last modifier are then used in rendering processes to generate the 3D representation.

Abstract: A caching system for a 3D modeling system. Intermediate channel results, created during the generation of a representation of an object, are cached. The caching increases the speed of system. In particular, the representation of the object is generated in multiple channels, each channel representing some portion of that representation. Intermediate channel results are generated in each channel by elements in that object's list of modifiers. Depending on for how long these intermediate channel results are valid, the intermediate channel results may or may not be cached. For example, if it is determined that one set of intermediate channel results will be valid for a predetermined period of time, while the next intermediate channel results in that channel will not be valid for a predetermined period of time, then the intermediate channel results are cached.

Abstract: A graphical user interface ("GUI") incorporating one or more subpanels. Each subpanel has a presentation control and can have one or more GUI objects displayed on the computer screen. Each subpanel may be toggled between an expanded state or a collapsed state by operating the presentation control. Expanding and collapsing the subpanels changes the appearance and the functionality of the GUI, but the size of area of the screen used by the subpanels remains unchanged.

Abstract: A method and system for controlling parameter values. One such system includes a number of functions for controlling parameter values, i.e. "parameter control functions," to enable a user to quickly and easily control variable parameter values using a graphical user interface and to enable the user to obtain the desired precision of control of such variables. One such system provides four control functions: a data entry function, single step function, a scroll function, and a translation function. The user can select the particular control function used to control a parameter value in light of the particular amount or type of control that needs to be accomplished.