Nvidia Corporation

Rajath Shetty

Niranjan Avadhanam

Nishant Puri

Shagan Sah

Hairong Jiang

Nuri Murat Arar

Pavlo Molchanov

Yuzhuo Ren

Sakthivel Sivaraman

Sujay Yadawadkar

Manoj Kumar Yennapureddy

Arjun Guru

Nuri Arar Murat

In various examples, occupant assessment using multi-modal sensor fusion for monitoring systems and applications are provided. In some embodiments, an occupant monitoring system comprises an occupant evaluation function that may predict at least one characteristic representative of a size of the occupant. The occupant evaluation function may include a first processing path that generates a representation of features corresponding to the occupant based on optical image data, and a second processing path that performs operations to determine a depth corresponding to the one or more features based on depth data derived from the optical image data and the point cloud depth data. In some embodiments, a three-dimensional pose detection model generates a three-dimensional pose estimate of the occupant using the optical image data, and the three-dimensional pose estimate is scaled to an absolute pose based on the point cloud depth data.

Occupant evaluation using multi-modal sensor fusion for in-cabin monitoring systems and applications

In various examples, techniques are described for detecting whether spoofing attacks are occurring using multiple sensors. Systems and methods are disclosed that include at least a first sensor having a first pose to capture a first perspective view of a user and a second sensor having a second pose to capture a second perspective view of the user. The first sensor and/or the second sensor may include an image sensor, a depth sensor, and/or the like. The systems and methods include a neural network that is configured to analyze first sensor data generated by the first sensor and second sensor data generated by the second sensor to determine whether a spoofing attack is occurring. The systems and methods may also perform one or more processes, such as facial recognition, based on whether the spoofing attack is occurring.

Detecting identity spoofing attacks in multi-sensor systems and applications

Systems and methods for determining the gaze direction of a subject and projecting this gaze direction onto specific regions of an arbitrary three-dimensional geometry. In an exemplary embodiment, gaze direction may be determined by a regression-based machine learning model. The determined gaze direction is then projected onto a three-dimensional map or set of surfaces that may represent any desired object or system. Maps may represent any three-dimensional layout or geometry, whether actual or virtual. Gaze vectors can thus be used to determine the object of gaze within any environment. Systems can also readily and efficiently adapt for use in different environments by retrieving a different set of surfaces or regions for each environment.

Neural network based determination of gaze direction using spatial models

A neural network system leverages dual attention, specifically both spatial attention and channel attention, to jointly estimate heart rate and respiratory rate of a subject by processing images of the subject. A motion neural network receives images of the subject and estimates heart and breath rates of the subject using both spatial and channel domain attention masks to focus processing on particular feature data. An appearance neural network computes a spatial attention mask from the images of the subject and may indicate that features associated with the subject's face (as opposed to the subject's hair or shoulders) to accurately estimate the heart and/or breath rate. Channel-wise domain attention is learned during training and recalibrates channel-wise feature responses to select the most informative features for processing. The channel attention mask is learned during training and can be used for different subjects during deployment.

Joint estimation of heart rate and respiratory rate using neural networks

Systems and methods for more accurate and robust determination of subject characteristics from an image of the subject. One or more machine learning models receive as input an image of a subject, and output both facial landmarks and associated confidence values. Confidence values represent the degrees to which portions of the subject's face corresponding to those landmarks are occluded, i.e., the amount of uncertainty in the position of each landmark location. These landmark points and their associated confidence values, and/or associated information, may then be input to another set of one or more machine learning models which may output any facial analysis quantity or quantities, such as the subject's gaze direction, head pose, drowsiness state, cognitive load, or distraction state.

Neural network based facial analysis using facial landmarks and associated confidence values

In various examples, an adaptive eye tracking machine learning model engine ('adaptive-model engine') for an eye tracking system is described. The adaptive-model engine may include an eye tracking or gaze tracking development pipeline ('adaptive-model training pipeline') that supports collecting data, training, optimizing, and deploying an adaptive eye tracking model that is a customized eye tracking model based on a set of features of an identified deployment environment. The adaptive-model engine supports ensembling the adaptive eye tracking model that may be trained on gaze vector estimation in surround environments and ensemble based on a plurality of eye tracking variant models and a plurality of facial landmark neural network metrics.

Adaptive eye tracking machine learning model engine

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.

Growing community of inventors

Santa Clara, CA, United States of America

Rajath Shetty