公开(公告)号：US20190244331A1

公开(公告)日：2019-08-08

申请号：US15887907

申请日：2018-02-02

Applicant: NVIDIA Corp.

Inventor： Jinwei Gu ,  Orazio Gallo ,  Ming-Yu Liu ,  Jan Kautz ,  Huaijin Chen

IPC: G06T5/00 ,  G06K9/46 ,  G06K9/66 ,  G06N3/08

Abstract: An image processing method extracts consecutive input blurry frames from a video, and generates sharp frames corresponding to the input blurry frames. An optical flow is determined between the sharp frames, and the optical flow is used to compute a per-pixel blur kernel. The blur kernel is used to reblur each of the sharp frames into a corresponding re-blurred frame. The re-blurred frame is used to fine-tune the deblur network by minimizing the distance between the re-blurred frame and the input blurry frame.

公开(公告)号：US20190340728A1

公开(公告)日：2019-11-07

申请号：US16130871

申请日：2018-09-13

Applicant: NVIDIA Corp.

Inventor： Varun Jampani ,  Deqing Sun ,  Ming-Yu Liu ,  Jan Kautz

IPC: G06T3/40 ,  G06T7/10 ,  G06N3/08 ,  G06K9/62

Abstract: A superpixel sampling network utilizes a neural network coupled to a differentiable simple linear iterative clustering component to determine pixel-superpixel associations from a set of pixel features output by the neural network. The superpixel sampling network computes updated superpixel centers and final pixel-superpixel associations over a number of iterations.

公开(公告)号：US20220261650A1

公开(公告)日：2022-08-18

申请号：US17346100

申请日：2021-06-11

Applicant: NVIDIA Corp.

Inventor： Jiawei Zhao ,  Steve Haihang Dai ,  Rangharajan Venkatesan ,  Ming-Yu Liu ,  William James Dally ,  Anima Anandkumar

IPC: G06N3/08 ,  G06F7/483 ,  G06F7/487 ,  G06F5/01 ,  G06F1/03

Abstract: An end-to-end low-precision training system based on a multi-base logarithmic number system and a multiplicative weight update algorithm. The multi-base logarithmic number system is applied to update weights of the neural network, with different bases of the multi-base logarithmic number system utilized between calculation of weight updates, calculation of feed-forward signals, and calculation of feedback signals. The LNS expresses a high dynamic range and computational energy efficiency, making it advantageous for on-board training in energy-constrained edge devices.

公开(公告)号：US10789678B2

公开(公告)日：2020-09-29

申请号：US16130871

申请日：2018-09-13

Applicant: NVIDIA Corp.

Inventor： Varun Jampani ,  Deqing Sun ,  Ming-Yu Liu ,  Jan Kautz

IPC: G06T3/40 ,  G06T7/10 ,  G06K9/62 ,  G06N3/08

Abstract: A superpixel sampling network utilizes a neural network coupled to a differentiable simple linear iterative clustering component to determine pixel-superpixel associations from a set of pixel features output by the neural network. The superpixel sampling network computes updated superpixel centers and final pixel-superpixel associations over a number of iterations.

公开(公告)号：US20230161467A1

公开(公告)日：2023-05-25

申请号：US17954872

申请日：2022-09-28

Applicant: NVIDIA Corp.

Inventor： Xun Huang ,  Ting Wang ,  Arun Mallya ,  Ming-Yu Liu

IPC: G06F3/04845 ,  G06T11/60 ,  G06F3/04847

CPC classification number: G06F3/04845 ,  G06T11/60 ,  G06F3/04847

Abstract: Techniques are described for synthesizing images with multiple input modalities. According to some embodiments, a user interface is displayed. The user interface comprises an input area operable to display one or more inputs corresponding to one or more input modalities, an output area operable to display an output image generated based on the one or more inputs, and at least one control operable to select which ones of the one or more input modalities should influence generation of the output image.

公开(公告)号：US10593020B2

公开(公告)日：2020-03-17

申请号：US15887907

申请日：2018-02-02

Applicant: NVIDIA Corp.

Inventor： Jinwei Gu ,  Orazio Gallo ,  Ming-Yu Liu ,  Jan Kautz ,  Huaijin Chen

IPC: G06K9/00 ,  G06T5/00 ,  G06N3/08 ,  G06K9/66 ,  G06K9/46 ,  G06K9/62

Abstract: An image processing method extracts consecutive input blurry frames from a video, and generates sharp frames corresponding to the input blurry frames. An optical flow is determined between the sharp frames, and the optical flow is used to compute a per-pixel blur kernel. The blur kernel is used to reblur each of the sharp frames into a corresponding re-blurred frame. The re-blurred frame is used to fine-tune the deblur network by minimizing the distance between the re-blurred frame and the input blurry frame.