-
公开(公告)号：US11056069B2
公开(公告)日：2021-07-06
申请号：US15953887
申请日：2018-04-16
Applicant: Google LLC
Inventor： Jonathan Huang , Gaurav Shah
Abstract: A method for driving a liquid crystal display (LCD) panel includes sequentially buffering each row of pixel data of a first display image in a corresponding pixel row of the LCD panel. The method also includes activating a backlight of the LCD panel after the last row of pixel data of the first display image has been buffered at the last pixel row of the LCD panel but before liquid crystal settling of the last pixel row of the LCD panel has completed. The method also may include initiating sequential buffering of each row of pixel data of a second display image in a corresponding pixel row of the LCD panel prior to the liquid crystal settling of the last pixel row of the LCD panel completing, wherein activating the backlight of the LCD panel comprises activating the backlight while at least one pixel row of the LCD panel buffers a corresponding row of pixel data from the second display image and other pixel rows of the LCD panel buffer corresponding rows of pixel data from the first display image.
-
公开(公告)号：US20230419538A1
公开(公告)日：2023-12-28
申请号：US18464912
申请日：2023-09-11
Applicant: Google LLC
Inventor： Yinxiao Li , Zhichao Lu , Xuehan Xiong , Jonathan Huang
IPC: G06T7/73
CPC classification number: G06T7/73 , G06T2207/20081 , G06T2207/30196 , G06T2207/20084 , G06T2207/10016
Abstract: A method includes receiving video data that includes a series of frames of image data. Here, the video data is representative of an actor performing an activity. The method also includes processing the video data to generate a spatial input stream including a series of spatial images representative of spatial features of the actor performing the activity, a temporal input stream representative of motion of the actor performing the activity, and a pose input stream including a series of images representative of a pose of the actor performing the activity. Using at least one neural network, the method also includes processing the temporal input stream, the spatial input stream, and the pose input stream. The method also includes classifying, by the at least one neural network, the activity based on the temporal input stream, the spatial input stream, and the pose input stream.
-
公开(公告)号：US20210334624A1
公开(公告)日：2021-10-28
申请号：US17365939
申请日：2021-07-01
Applicant: Google LLC
Inventor： Wei Hua , Barret Zoph , Jonathon Shlens , Chenxi Liu , Jonathan Huang , Jia Li , Fei-Fei Li , Kevin Patrick Murphy
Abstract: A method for determining an architecture for a task neural network configured to perform a particular machine learning task is described. The method includes obtaining data specifying a current set of candidate architectures for the task neural network; for each candidate architecture in the current set: processing the data specifying the candidate architecture using a performance prediction neural network having multiple performance prediction parameters, the performance prediction neural network being configured to process the data specifying the candidate architecture in accordance with current values of the performance prediction parameters to generate a performance prediction that characterizes how well a neural network having the candidate architecture would perform after being trained on the particular machine learning task; and generating an updated set of candidate architectures by selecting one or more of the candidate architectures in the current set based on the performance predictions for the candidate architectures in the current set.
-
公开(公告)号：US20200257961A1
公开(公告)日：2020-08-13
申请号：US16861491
申请日：2020-04-29
Applicant: Google LLC
Inventor： Wei Hua , Barret Zoph , Jonathon Shlens , Chenxi Liu , Jonathan Huang , Jia Li , Fei-Fei Li , Kevin Patrick Murphy
Abstract: A method for determining an architecture for a task neural network configured to perform a particular machine learning task is described. The method includes obtaining data specifying a current set of candidate architectures for the task neural network; for each candidate architecture in the current set: processing the data specifying the candidate architecture using a performance prediction neural network having multiple performance prediction parameters, the performance prediction neural network being configured to process the data specifying the candidate architecture in accordance with current values of the performance prediction parameters to generate a performance prediction that characterizes how well a neural network having the candidate architecture would perform after being trained on the particular machine learning task; and generating an updated set of candidate architectures by selecting one or more of the candidate architectures in the current set based on the performance predictions for the candidate architectures in the current set.
-
公开(公告)号：US10116922B2
公开(公告)日：2018-10-30
申请号：US15289008
申请日：2016-10-07
Applicant: Google LLC
Inventor： Jonathan Huang , Samuel Kvaalen , Peter Bradshaw
IPC: G06T15/00 , H04N13/00 , H04N13/20 , H04N13/156 , H04N13/221 , G06T7/00 , H04N13/02 , G03B35/02 , G06K9/00 , G06T15/20 , G06T7/593 , G06T7/285 , G06K9/62 , G06K9/66 , G06T7/70
Abstract: Disclosed herein are methods, devices, and non-transitory computer readable media that relate to stereoscopic image creation. A camera captures an initial image at an initial position. A target displacement from the initial position is determined for a desired stereoscopic effect, and an instruction is provided that specifies a direction in which to move the camera from the initial position. While the camera is in motion, an estimated displacement from the initial position is calculated. When the estimated displacement corresponds to the target displacement, the camera automatically captures a candidate image. An acceptability analysis is performed to determine whether the candidate image has acceptable image quality and acceptable similarity to the initial image. If the candidate image passes the acceptability analysis, a stereoscopic image is created based on the initial and candidate images.
-
Patent Agency Ranking
公开(公告)号：US20180176780A1
公开(公告)日：2018-06-21
申请号：US15847121
申请日：2017-12-19
Applicant: Google LLC
Inventor： Jonathan Huang
CPC classification number: H04W12/06 , G06F1/26 , H04L41/0809 , H04L41/0886 , H04L63/0846 , H04L63/102 , H04W12/04 , H04W12/08
Abstract: Methods, systems, and apparatuses, including computer programs encoded on a computer-readable storage medium for implementing network connectivity using power lines are described. After a device is connected to a power line, wireless network credentials that enable the device to access to a wireless network are transmitted to the device through the power line. A request is received from the device to connect to the wireless network. The request includes the wireless network credentials transmitted to the device through the power line. The device is determined to be authorized to access the wireless network based on the one or more credentials received in the request. In response to determining that the one or more wireless network credentials received in the request from the device are valid, wireless network access is provided to the device.
-
公开(公告)号：US11776156B2
公开(公告)日：2023-10-03
申请号：US17303969
申请日：2021-06-11
Applicant: Google LLC
Inventor： Yinxiao Li , Zhichao Lu , Xuehan Xiong , Jonathan Huang
IPC: G06T7/73
CPC classification number: G06T7/73 , G06T2207/10016 , G06T2207/20081 , G06T2207/20084 , G06T2207/30196
Abstract: A method includes receiving video data that includes a series of frames of image data. Here, the video data is representative of an actor performing an activity. The method also includes processing the video data to generate a spatial input stream including a series of spatial images representative of spatial features of the actor performing the activity, a temporal input stream representative of motion of the actor performing the activity, and a pose input stream including a series of images representative of a pose of the actor performing the activity. Using at least one neural network, the method also includes processing the temporal input stream, the spatial input stream, and the pose input stream. The method also includes classifying, by the at least one neural network, the activity based on the temporal input stream, the spatial input stream, and the pose input stream.
-
公开(公告)号：US20210390733A1
公开(公告)日：2021-12-16
申请号：US17303969
申请日：2021-06-11
Applicant: Google LLC
Inventor： Yinxiao Li , Zhichao Lu , Xuehan Xiong , Jonathan Huang
IPC: G06T7/73
Abstract: A method includes receiving video data that includes a series of frames of image data. Here, the video data is representative of an actor performing an activity. The method also includes processing the video data to generate a spatial input stream including a series of spatial images representative of spatial features of the actor performing the activity, a temporal input stream representative of motion of the actor performing the activity, and a pose input stream including a series of images representative of a pose of the actor performing the activity. Using at least one neural network, the method also includes processing the temporal input stream, the spatial input stream, and the pose input stream. The method also includes classifying, by the at least one neural network, the activity based on the temporal input stream, the spatial input stream, and the pose input stream.
-
公开(公告)号：US11089643B2
公开(公告)日：2021-08-10
申请号：US15984235
申请日：2018-05-18
Applicant: Google LLC
Inventor： Jonathan Huang , Paul Stewart
IPC: H04W76/14 , H04W88/04 , H04W4/80 , H04L12/24 , H04L12/26 , H04W76/30 , H04W84/18 , H04W48/18 , H04W24/10 , H04W24/00 , H04W12/06 , H04W84/20 , H04W88/06 , H04W12/062 , H04W48/12
Abstract: A device that implements adaptive on-demand tethering may include at least one processor circuit. The at least one processor circuit may be configured to monitor at least a first connection quality value associated with a first network connection of the device to a network. The at least one processor circuit may be further configured to receive information regarding a second connection quality value associated with a second network connection of another device. The at least one processor circuit may be further configured to initiate a tethering connection with the another device based at least in part on a comparison of the first connection quality value and the second connection quality value. The at least one processor circuit may be further configured to connect to the network through the second network connection of the another device via the tethering connection based at least in part on the comparison.
-
公开(公告)号：US11087201B2
公开(公告)日：2021-08-10
申请号：US16861491
申请日：2020-04-29
Applicant: Google LLC
Inventor： Wei Hua , Barret Zoph , Jonathon Shlens , Chenxi Liu , Jonathan Huang , Jia Li , Fei-Fei Li , Kevin Patrick Murphy
Abstract: A method for determining an architecture for a task neural network configured to perform a particular machine learning task is described. The method includes obtaining data specifying a current set of candidate architectures for the task neural network; for each candidate architecture in the current set: processing the data specifying the candidate architecture using a performance prediction neural network having multiple performance prediction parameters, the performance prediction neural network being configured to process the data specifying the candidate architecture in accordance with current values of the performance prediction parameters to generate a performance prediction that characterizes how well a neural network having the candidate architecture would perform after being trained on the particular machine learning task; and generating an updated set of candidate architectures by selecting one or more of the candidate architectures in the current set based on the performance predictions for the candidate architectures in the current set.
-
-
-
-
-
-
-
-
-
Search Results
13
records
(took 0.015 seconds)
