公开(公告)号：US20230326182A1

公开(公告)日：2023-10-12

申请号：US18333281

申请日：2023-06-12

Applicant: NVIDIA Corporation

Inventor： Sakthivel Sivaraman ,  Shagan Sah ,  Niranjan Avadhanam

IPC: G06V10/764 ,  G06V20/40 ,  G06V40/20 ,  G06V40/10 ,  G06F18/20 ,  G06F18/2321 ,  G06N3/045

CPC classification number: G06V10/764 ,  G06V20/47 ,  G06V40/28 ,  G06V40/113 ,  G06F18/285 ,  G06F18/2321 ,  G06N3/045

Abstract: In various examples, the present disclosure relates to using temporal filters for automated real-time classification. The technology described herein improves the performance of a multiclass classifier that may be used to classify a temporal sequence of input signals—such as input signals representative of video frames. A performance improvement may be achieved, at least in part, by applying a temporal filter to an output of the multiclass classifier. For example, the temporal filter may leverage classifications associated with preceding input signals to improve the final classification given to a subsequent signal. In some embodiments, the temporal filter may also use data from a confusion matrix to correct for the probable occurrence of certain types of classification errors. The temporal filter may be a linear filter, a nonlinear filter, an adaptive filter, and/or a statistical filter.

公开(公告)号：US20220410830A1

公开(公告)日：2022-12-29

申请号：US17939613

申请日：2022-09-07

Applicant: NVIDIA Corporation

Inventor： Atousa Torabi ,  Sakthivel Sivaraman ,  Niranjan Avadhanam ,  Shagan Sah

IPC: B60R21/017 ,  B60R21/013 ,  B60W60/00 ,  G06N3/02 ,  B60W50/14

Abstract: In various examples, systems and methods are disclosed that accurately identify driver and passenger in-cabin activities that may indicate a biomechanical distraction that prevents a driver from being fully engaged in driving a vehicle. In particular, image data representative of an image of an occupant of a vehicle may be applied to one or more deep neural networks (DNNs). Using the DNNs, data indicative of key point locations corresponding to the occupant may be computed, a shape and/or a volume corresponding to the occupant may be reconstructed, a position and size of the occupant may be estimated, hand gesture activities may be classified, and/or body postures or poses may be classified. These determinations may be used to determine operations or settings for the vehicle to increase not only the safety of the occupants, but also of surrounding motorists, bicyclists, and pedestrians.

公开(公告)号：US11144754B2

公开(公告)日：2021-10-12

申请号：US16544442

申请日：2019-08-19

Applicant: Nvidia Corporation

Inventor： Feng Hu ,  Niranjan Avadhanam ,  Yuzhuo Ren ,  Sujay Yadawadkar ,  Sakthivel Sivaraman ,  Hairong Jiang ,  Siyue Wu

IPC: G06K9/00 ,  G06N3/08 ,  G06F3/01 ,  G06K9/62

Abstract: Apparatuses, systems, and techniques are described to determine locations of objects using images including digital representations of those objects. In at least one embodiment, a gaze of one or more occupants of a vehicle is determined independently of a location of one or more sensors used to detect those occupants.

公开(公告)号：US12236351B2

公开(公告)日：2025-02-25

申请号：US18497501

申请日：2023-10-30

Applicant: Nvidia Corporation

Inventor： Feng Hu ,  Niranjan Avadhanam ,  Yuzhuo Ren ,  Sujay Yadawadkar ,  Sakthivel Sivaraman ,  Hairong Jiang ,  Siyue Wu

IPC: G06N3/084 ,  G06F3/01 ,  G06F18/21 ,  G06N3/08 ,  G06V10/764 ,  G06V10/82 ,  G06V20/59 ,  G06V40/18 ,  G06V40/19

Abstract: Apparatuses, systems, and techniques are described to determine locations of objects using images including digital representations of those objects. In at least one embodiment, a gaze of one or more occupants of a vehicle is determined independently of a location of one or more sensors used to detect those occupants.

公开(公告)号：US11721089B2

公开(公告)日：2023-08-08

申请号：US17647390

申请日：2022-01-07

Applicant: NVIDIA Corporation

Inventor： Sakthivel Sivaraman ,  Shagan Sah ,  Niranjan Avadhanam

IPC: G06V10/00 ,  G06V10/764 ,  G06V20/40 ,  G06V40/20 ,  G06V40/10 ,  G06F18/20 ,  G06F18/2321 ,  G06N3/045

CPC classification number: G06V10/764 ,  G06F18/2321 ,  G06F18/285 ,  G06N3/045 ,  G06V20/47 ,  G06V40/113 ,  G06V40/28

Abstract: In various examples, the present disclosure relates to using temporal filters for automated real-time classification. The technology described herein improves the performance of a multiclass classifier that may be used to classify a temporal sequence of input signals—such as input signals representative of video frames. A performance improvement may be achieved, at least in part, by applying a temporal filter to an output of the multiclass classifier. For example, the temporal filter may leverage classifications associated with preceding input signals to improve the final classification given to a subsequent signal. In some embodiments, the temporal filter may also use data from a confusion matrix to correct for the probable occurrence of certain types of classification errors. The temporal filter may be a linear filter, a nonlinear filter, an adaptive filter, and/or a statistical filter.

公开(公告)号：US20230064049A1

公开(公告)日：2023-03-02

申请号：US17462833

申请日：2021-08-31

Applicant: Nvidia Corporation

Inventor： Sakthivel Sivaraman ,  Nishant Puri ,  Yuzhuo Ren ,  Atousa Torabi ,  Shubhadeep Das ,  Niranjan Avadhanam ,  Sumit Kumar Bhattacharya ,  Jason Roche

IPC: G06K9/00 ,  G06T7/73 ,  G06F3/01 ,  G06F16/632 ,  G06T15/06

Abstract: Interactions with virtual systems may be difficult when users inadvertently fail to provide sufficient information to proceed with their requests. Certain types of inputs, such as auditory inputs, may lack sufficient information to properly provide a response to the user. Additional information, such as image data, may enable user gestures or poses to supplement the auditory inputs to enable response generation without requesting additional information from users.

公开(公告)号：US11222232B1

公开(公告)日：2022-01-11

申请号：US16907125

申请日：2020-06-19

Applicant: NVIDIA Corporation

Inventor： Sakthivel Sivaraman ,  Shagan Sah ,  Niranjan Avadhanam

IPC: G06K9/00 ,  G06K9/62 ,  G06N3/04 ,  G06K9/42

Abstract: In various examples, the present disclosure relates to using temporal filters for automated real-time classification. The technology described herein improves the performance of a multiclass classifier that may be used to classify a temporal sequence of input signals—such as input signals representative of video frames. A performance improvement may be achieved, at least in part, by applying a temporal filter to an output of the multiclass classifier. For example, the temporal filter may leverage classifications associated with preceding input signals to improve the final classification given to a subsequent signal. In some embodiments, the temporal filter may also use data from a confusion matrix to correct for the probable occurrence of certain types of classification errors. The temporal filter may be a linear filter, a nonlinear filter, an adaptive filter, and/or a statistical filter.

公开(公告)号：US20210397885A1

公开(公告)日：2021-12-23

申请号：US16907125

申请日：2020-06-19

Applicant: NVIDIA Corporation

Inventor： Sakthivel Sivaraman ,  Shagan Sah ,  Niranjan Avadhanam

IPC: G06K9/62 ,  G06K9/00 ,  G06N3/04 ,  G06K9/42

Abstract: In various examples, the present disclosure relates to using temporal filters for automated real-time classification. The technology described herein improves the performance of a multiclass classifier that may be used to classify a temporal sequence of input signals—such as input signals representative of video frames. A performance improvement may be achieved, at least in part, by applying a temporal filter to an output of the multiclass classifier. For example, the temporal filter may leverage classifications associated with preceding input signals to improve the final classification given to a subsequent signal. In some embodiments, the temporal filter may also use data from a confusion matrix to correct for the probable occurrence of certain types of classification errors. The temporal filter may be a linear filter, a nonlinear filter, an adaptive filter, and/or a statistical filter.

公开(公告)号：US20250050831A1

公开(公告)日：2025-02-13

申请号：US18931478

申请日：2024-10-30

Applicant: NVIDIA Corporation

Inventor： Atousa Torabi ,  Sakthivel Sivaraman ,  Niranjan Avadhanam ,  Shagan Sah

IPC: B60R21/017 ,  B60R21/01 ,  B60R21/013 ,  B60W50/00 ,  B60W50/14 ,  B60W60/00 ,  G06N3/02

Abstract: In various examples, systems and methods are disclosed that accurately identify driver and passenger in-cabin activities that may indicate a biomechanical distraction that prevents a driver from being fully engaged in driving a vehicle. In particular, image data representative of an image of an occupant of a vehicle may be applied to one or more deep neural networks (DNNs). Using the DNNs, data indicative of key point locations corresponding to the occupant may be computed, a shape and/or a volume corresponding to the occupant may be reconstructed, a position and size of the occupant may be estimated, hand gesture activities may be classified, and/or body postures or poses may be classified. These determinations may be used to determine operations or settings for the vehicle to increase not only the safety of the occupants, but also of surrounding motorists, bicyclists, and pedestrians.

公开(公告)号：US11688074B2

公开(公告)日：2023-06-27

申请号：US17039437

申请日：2020-09-30

Applicant: NVIDIA Corporation

Inventor： Nishant Puri ,  Sakthivel Sivaraman ,  Rajath Shetty ,  Niranjan Avadhanam

IPC: G06T7/194 ,  G06T5/30 ,  G06T5/00 ,  G06N3/08 ,  G06V40/10 ,  G06F18/214 ,  G06F18/24

CPC classification number: G06T7/194 ,  G06F18/214 ,  G06F18/24 ,  G06N3/08 ,  G06T5/002 ,  G06T5/30 ,  G06V40/11 ,  G06V40/113 ,  G06T2207/20081 ,  G06T2207/20084 ,  G06T2207/20132 ,  G06T2207/20221 ,  G06T2207/30196

Abstract: In various examples, a background of an object may be modified to generate a training image. A segmentation mask may be generated and used to generate an object image that includes image data representing the object. The object image may be integrated into a different background and used for data augmentation in training a neural network. Data augmentation may also be performed using hue adjustment (e.g., of the object image) and/or rendering three-dimensional capture data that corresponds to the object from selected views. Inference scores may be analyzed to select a background for an image to be included in a training dataset. Backgrounds may be selected and training images may be added to a training dataset iteratively during training (e.g., between epochs). Additionally, early or late fusion nay be employed that uses object mask data to improve inferencing performed by a neural network trained using object mask data.