-
公开(公告)号：US20200184313A1
公开(公告)日：2020-06-11
申请号：US16050362
申请日：2018-07-31
申请人： Pixar , Disney Enterprises, Inc.
发明人： Thijs Vogels , Fabrice Rousselle , Jan Novak , Brian McWilliams , Mark Meyer , Alex Harvill
摘要： A modular architecture is provided for denoising Monte Carlo renderings using neural networks. The temporal approach extracts and combines feature representations from neighboring frames rather than building a temporal context using recurrent connections. A multiscale architecture includes separate single-frame or temporal denoising modules for individual scales, and one or more scale compositor neural networks configured to adaptively blend individual scales. An error-predicting module is configured to produce adaptive sampling maps for a renderer to achieve more uniform residual noise distribution. An asymmetric loss function may be used for training the neural networks, which can provide control over the variance-bias trade-off during denoising.
-
公开(公告)号：US20190304069A1
公开(公告)日：2019-10-03
申请号：US16050336
申请日：2018-07-31
申请人： Pixar , Disney Enterprises, Inc.
发明人： Thijs Vogels , Fabrice Rousselle , Jan Novak , Brian McWilliams , Mark Meyer , Alex Harvill , David Adler
摘要： A modular architecture is provided for denoising Monte Carlo renderings using neural networks. The temporal approach extracts and combines feature representations from neighboring frames rather than building a temporal context using recurrent connections. A multiscale architecture includes separate single-frame or temporal denoising modules for individual scales, and one or more scale compositor neural networks configured to adaptively blend individual scales. An error-predicting module is configured to produce adaptive sampling maps for a renderer to achieve more uniform residual noise distribution. An asymmetric loss function may be used for training the neural networks, which can provide control over the variance-bias trade-off during denoising.
-
公开(公告)号：US20190304068A1
公开(公告)日：2019-10-03
申请号：US16050332
申请日：2018-07-31
申请人： Pixar , Disney Enterprises, Inc.
发明人： Thijs Vogels , Fabrice Rousselle , Jan Novak , Brian McWilliams , Mark Meyer , Alex Harvill
摘要： A modular architecture is provided for denoising Monte Carlo renderings using neural networks. The temporal approach extracts and combines feature representations from neighboring frames rather than building a temporal context using recurrent connections. A multiscale architecture includes separate single-frame or temporal denoising modules for individual scales, and one or more scale compositor neural networks configured to adaptively blend individual scales. An error-predicting module is configured to produce adaptive sampling maps for a renderer to achieve more uniform residual noise distribution. An asymmetric loss function may be used for training the neural networks, which can provide control over the variance-bias trade-off during denoising.
-
公开(公告)号：US20230083929A1
公开(公告)日：2023-03-16
申请号：US17984132
申请日：2022-11-09
申请人： Pixar , Disney Enterprises, Inc.
发明人： Thijs Vogels , Fabrice Rousselle , Jan Novak , Brian McWilliams , Mark Meyer , Alex Harvill
IPC分类号： G06T5/00 , G06N3/08 , G06T15/06 , G06T15/50 , G06N5/04 , G06N20/00 , G06F17/18 , G06N3/04 , G06T5/50
摘要： A modular architecture is provided for denoising Monte Carlo renderings using neural networks. The temporal approach extracts and combines feature representations from neighboring frames rather than building a temporal context using recurrent connections. A multiscale architecture includes separate single-frame or temporal denoising modules for individual scales, and one or more scale compositor neural networks configured to adaptively blend individual scales. An error-predicting module is configured to produce adaptive sampling maps for a renderer to achieve more uniform residual noise distribution. An asymmetric loss function may be used for training the neural networks, which can provide control over the variance-bias trade-off during denoising.
-
公开(公告)号：US11532073B2
公开(公告)日：2022-12-20
申请号：US16050314
申请日：2018-07-31
申请人： Pixar , Disney Enterprises, Inc.
发明人： Thijs Vogels , Fabrice Rousselle , Jan Novak , Brian McWilliams , Mark Meyer , Alex Harvill
IPC分类号： G06T15/50 , G06T15/06 , G06T7/90 , G06T5/00 , G06T7/00 , G06N3/04 , G06N3/08 , G06N5/04 , G06N7/00 , G06N20/00 , G06F17/18 , G06T5/50
摘要： A modular architecture is provided for denoising Monte Carlo renderings using neural networks. The temporal approach extracts and combines feature representations from neighboring frames rather than building a temporal context using recurrent connections. A multiscale architecture includes separate single-frame or temporal denoising modules for individual scales, and one or more scale compositor neural networks configured to adaptively blend individual scales. An error-predicting module is configured to produce adaptive sampling maps for a renderer to achieve more uniform residual noise distribution. An asymmetric loss function may be used for training the neural networks, which can provide control over the variance-bias trade-off during denoising.
-
公开(公告)号：US11037274B2
公开(公告)日：2021-06-15
申请号：US16789025
申请日：2020-02-12
申请人： PIXAR , Disney Enterprises, Inc.
发明人： Thijs Vogels , Fabrice Rousselle , Brian McWilliams , Mark Meyer , Jan Novak
IPC分类号： G06K9/62 , G06T5/00 , G06T5/50 , G06N3/08 , G06N3/04 , G06K9/46 , G06T7/00 , G06T15/06 , G06T7/90
摘要： Supervised machine learning using neural networks is applied to denoising images rendered by MC path tracing. Specialization of neural networks may be achieved by using a modular design that allows reusing trained components in different networks and facilitates easy debugging and incremental building of complex structures. Specialization may also be achieved by using progressive neural networks. In some embodiments, training of a neural-network based denoiser may use importance sampling, where more challenging patches or patches including areas of particular interests within a training dataset are selected with higher probabilities than others. In some other embodiments, generative adversarial networks (GANs) may be used for training a machine-learning based denoiser as an alternative to using pre-defined loss functions.
-
公开(公告)号：US10789686B2
公开(公告)日：2020-09-29
申请号：US16735079
申请日：2020-01-06
申请人： PIXAR , Disney Enterprises, Inc.
发明人： Thijs Vogels , Fabrice Rousselle , Brian McWilliams , Mark Meyer , Jan Novak
IPC分类号： G06K9/62 , G06T5/00 , G06T5/50 , G06N3/08 , G06N3/04 , G06K9/46 , G06T7/00 , G06T15/06 , G06T7/90
摘要： Supervised machine learning using neural networks is applied to denoising images rendered by MC path tracing. Specialization of neural networks may be achieved by using a modular design that allows reusing trained components in different networks and facilitates easy debugging and incremental building of complex structures. Specialization may also be achieved by using progressive neural networks. In some embodiments, training of a neural-network based denoiser may use importance sampling, where more challenging patches or patches including areas of particular interests within a training dataset are selected with higher probabilities than others. In some other embodiments, generative adversarial networks (GANs) may be used for training a machine-learning based denoiser as an alternative to using pre-defined loss functions.
-
公开(公告)号：US10699382B2
公开(公告)日：2020-06-30
申请号：US16050336
申请日：2018-07-31
申请人： Pixar , Disney Enterprises, Inc.
发明人： Thijs Vogels , Fabrice Rousselle , Jan Novak , Brian McWilliams , Mark Meyer , Alex Harvill
IPC分类号： G06T5/00 , G06N3/08 , G06T15/06 , G06T15/50 , G06N5/04 , G06N20/00 , G06F17/18 , G06N3/04 , G06T5/50
摘要： A modular architecture is provided for denoising Monte Carlo renderings using neural networks. The temporal approach extracts and combines feature representations from neighboring frames rather than building a temporal context using recurrent connections. A multiscale architecture includes separate single-frame or temporal denoising modules for individual scales, and one or more scale compositor neural networks configured to adaptively blend individual scales. An error-predicting module is configured to produce adaptive sampling maps for a renderer to achieve more uniform residual noise distribution. An asymmetric loss function may be used for training the neural networks, which can provide control over the variance-bias trade-off during denoising.
-
公开(公告)号：US10607319B2
公开(公告)日：2020-03-31
申请号：US15946652
申请日：2018-04-05
申请人： PIXAR , Disney Enterprises, Inc.
发明人： Thijs Vogels , Fabrice Rousselle , Brian McWilliams , Mark Meyer , Jan Novak
IPC分类号： G06K9/62 , G06T5/00 , G06T5/50 , G06N3/08 , G06N3/04 , G06K9/46 , G06T7/00 , G06T15/06 , G06T7/90
摘要： Supervised machine learning using neural networks is applied to denoising images rendered by MC path tracing. Specialization of neural networks may be achieved by using a modular design that allows reusing trained components in different networks and facilitates easy debugging and incremental building of complex structures. Specialization may also be achieved by using progressive neural networks. In some embodiments, training of a neural-network based denoiser may use importance sampling, where more challenging patches or patches including areas of particular interests within a training dataset are selected with higher probabilities than others. In some other embodiments, generative adversarial networks (GANs) may be used for training a machine-learning based denoiser as an alternative to using pre-defined loss functions.
-
公开(公告)号：US10586310B2
公开(公告)日：2020-03-10
申请号：US15946649
申请日：2018-04-05
申请人： PIXAR , Disney Enterprises, Inc.
发明人： Thijs Vogels , Fabrice Rousselle , Brian McWilliams , Mark Meyer , Jan Novak
IPC分类号： G06K9/62 , G06T5/00 , G06T5/50 , G06N3/08 , G06N3/04 , G06K9/46 , G06T7/00 , G06T15/06 , G06T7/90
摘要： Supervised machine learning using neural networks is applied to denoising images rendered by MC path tracing. Specialization of neural networks may be achieved by using a modular design that allows reusing trained components in different networks and facilitates easy debugging and incremental building of complex structures. Specialization may also be achieved by using progressive neural networks. In some embodiments, training of a neural-network based denoiser may use importance sampling, where more challenging patches or patches including areas of particular interests within a training dataset are selected with higher probabilities than others. In some other embodiments, generative adversarial networks (GANs) may be used for training a machine-learning based denoiser as an alternative to using pre-defined loss functions.
-
-
-
-
-
-
-
-
-
搜索结果
23 条记录 (耗时 0.02 秒)
