Abstract: A task-inconsistency based data collection trigger method includes: receiving a plurality of outputs for a plurality of tasks from an advanced driving system (ADS), the plurality of tasks including a task to be verified and remaining tasks; determining a consistency estimate of one or more outputs of the task to be verified based on outputs of the remaining tasks; and generating a data collection trigger based on the consistency estimate being below a consistency threshold. The method may further include collecting data based on the data collection trigger and adding the collected data to one or more data sets for training the ADS.

Abstract: This disclosure provides systems, methods, and devices that support machine learning techniques for vehicle guidance. In one aspect, a method is provided that includes receiving an image frame from an image sensor, determining a projected image frame to correct a vertical axis of the image frame, and generating a top-down representation of the environment based on the projected image frame. The projected image frame may be aligned with the top down representation's vertical axis and may be projected using a cylindrical family projection. The top-down representation may be produced by assembling depth and context matrices from vertically aligned pixels in the projected image frame.

Abstract: Systems and techniques are described herein for sensor feature projection. For example, a computing device can generate, using an encoder, a first plurality of image features from a first plurality of images of an environment; process the first plurality of image features to generate a first plurality of aerial view features, wherein each aerial view feature from the first plurality of aerial view features is associated with a respective region of the environment; generate a first mask associated with a first region of the environment and aerial view features of the first plurality of aerial view features; generate, using the encoder, a second plurality of image features from a second plurality of images of the environment; and process, using the first mask, one or more image features of the second plurality of image features to generate a second plurality of aerial view features associated with the first region.

Abstract: An apparatus configured to perform a perception task may generate sensor features from data from one or more sensors. process the sensor features with a time-continuous recurrent neural network (RNN) to produce time-continuous features, and perform the perception task using the time-continuous features. The time-continuous features may be defined by a first feature vector value corresponding to a first observation time of the one or more sensors, a prediction of a steady state feature vector value, and estimated feature vector values between the first feature vector value and the steady state feature vector value, the estimated feature vector values being defined by a function.

Abstract: Systems and techniques are described herein for generating map data. For instance, a method for generating map data is provided. The method may include processing sensor data representative of a scene using a bird's-eye-view (BEV) detector to generate a BEV map of the scene; processing the BEV map using a BEV spatial prior model to generate one or more priors; and refining the BEV map, based on the one or more priors, to generate a refined BEV map.

Abstract: Various aspects of the present disclosure generally relate to autonomous driving and driver assistance systems. In some aspects, a device associated with an ego vehicle may obtain a series of images that depict a reference vehicle traveling along a road segment ahead of the ego vehicle. The device may estimate, based on the series of images, a size of the reference vehicle and/or a position of the reference vehicle relative to the ego vehicle. The device may track a trajectory of the reference vehicle along the road segment ahead of the ego vehicle based on the estimated size of the reference vehicle and/or the estimated position of the reference vehicle over the series of images. The device may estimate a surface geometry associated with the road segment ahead of the ego vehicle based on the tracked trajectory of the reference vehicle. Numerous other aspects are described.

Abstract: Systems and techniques are described herein for object detection. For example, a computing device can extract, by an encoder of the computing device, a plurality of features from one or more images of an environment of the computing device. The computing device can determine, based on the plurality of features, a first detection of one or more objects and three-dimensional (3D) coordinates for the one or more objects. The computing device can back-project the 3D coordinates of the one or more objects onto the one or more images. The computing device can determine one or more regions of at least one first image of the one or more images based on the back-projection of the 3D coordinates of the one or more objects. computing device determine, based on the one or more regions of the at least one first image, a second detection of the one or more objects.

Abstract: An apparatus comprises one or more memories, and one or more processors in communication with the one or more memories. The one or processors are configured to execute an automated driving system having a modular hybrid architecture. The modular hybrid architecture includes a plurality of task units, and the modular hybrid architecture includes one or more human-defined interfaces, and one or more AI-defined interfaces. The one or more processors are further configured to receive input data from one or more sensors process the input data using the automated driving system having the modular hybrid architecture, and control at least one operation of a vehicle according to an output of the automated driving system.

Abstract: Systems and techniques are described herein for tracking objects. For instance, a method for tracking objects is provided. The method may include generating an output embedding based on an object in a first image; obtaining a predicted position of the object in a second image; modifying the output embedding based on the predicted position to generate a modified output embedding; and detecting the object in the second image based on the modified output embedding.

Abstract: This disclosure provides systems, methods, and devices for vehicle driving assistance systems that support image processing for vehicular monitoring operations. In a first aspect, a method of monitoring includes receiving first image data from a first camera oriented in a first direction with a first field of view facing a user; receiving second image data from a second camera oriented in a second direction different from the first direction, the second camera having a second field of view corresponding to a field of view of the user; determining a set of regions of interest based on the second image data; determining a gaze direction of the user based on the first image data; and determining an attentiveness score based on correspondence between the set of regions of interest and the gaze direction. Other aspects and features are also claimed and described.

Abstract: A headlight control system (10) for a motor vehicle including, a controllable headlight (24) adapted to generate variable illumination of the vehicle environment, an imaging apparatus (11) adapted to capture images (100) from a region in front of the motor vehicle, and a data processing device (14) adapted to perform image processing of images (100) captured by the imaging apparatus (11) and to vary the light characteristics of the controllable headlight (24) depending on the image processing. A machine learning model (27) is implemented in the data processing device (14) which is trained to estimate and output an output signal (29) representing a desired illumination of the vehicle environment from one or more images (28) received as input from the imaging apparatus (11).

Abstract: A method of headlight evaluation can include receiving from a camera mounted on a vehicle, one or more images that include at least a portion of an illumination pattern projected by a headlight of the vehicle upon a surface of an object. The object can be a stationary object or a moving object such as, for example, a traffic sign or another vehicle. The surface can be a front portion of the traffic sign or a rear vertical surface of the other vehicle. The illumination pattern may be evaluated based on a positional relationship of the camera, the headlight, and the surface of the object. A headlight position status and/or a headlight orientation status can be output based on evaluating the illumination pattern.

Abstract: Various embodiments include methods and vehicle processing systems implementing such methods for protecting vehicles from dangers caused by animals on or near the roadway. In various embodiments, a vehicle processing system may perform a recognition process to identify an animal detected in proximity to the vehicle, perform a plurality of simulations of outcomes for the vehicle and other vehicles resulting from stimulating the identified animal using multiple different stimuli modes of the vehicle signal devices in which each different stimuli mode is predicted to elicit different animal behaviors, select one of the different stimulus modes to be performed by vehicle signal devices to elicit a behavior of the identified animal based on the plurality of simulated outcomes for the vehicle and other vehicles, and control signal devices of the vehicle to perform the selected stimulus.

Abstract: This disclosure provides systems, methods, and devices that support machine learning techniques for vehicle guidance. In one aspect, a method is provided that includes receiving an image frame from an image sensor, determining a projected image frame to correct a vertical axis of the image frame, and generating a top-down representation of the environment based on the projected image frame. The projected image frame may be aligned with the top down representation's vertical axis and may be projected using a cylindrical family projection. The top-down representation may be produced by assembling depth and context matrices from vertically aligned pixels in the projected image frame.

Abstract: Various aspects of the present disclosure generally relate to autonomous driving and driver assistance systems. In some aspects, a device associated with an ego vehicle may obtain a series of images that depict a reference vehicle traveling along a road segment ahead of the ego vehicle. The device may estimate, based on the series of images, a size of the reference vehicle and/or a position of the reference vehicle relative to the ego vehicle. The device may track a trajectory of the reference vehicle along the road segment ahead of the ego vehicle based on the estimated size of the reference vehicle and/or the estimated position of the reference vehicle over the series of images. The device may estimate a surface geometry associated with the road segment ahead of the ego vehicle based on the tracked trajectory of the reference vehicle. Numerous other aspects are described.

Abstract: This disclosure provides systems, methods, and devices for vehicle driving assistance systems that support image processing for vehicular monitoring operations. In a first aspect, a method of monitoring includes receiving first image data from a first camera oriented in a first direction with a first field of view facing a user; receiving second image data from a second camera oriented in a second direction different from the first direction, the second camera having a second field of view corresponding to a field of view of the user; determining a set of regions of interest based on the second image data; determining a gaze direction of the user based on the first image data; and determining an attentiveness score based on correspondence between the set of regions of interest and the gaze direction. Other aspects and features are also claimed and described.

Abstract: A vision system for a motor vehicle comprises an imaging apparatus (11) adapted to capture images (30) from a surrounding of the motor vehicle, and a data processing unit (14) adapted to perform image processing on images (30) captured by said imaging apparatus (11). The data processing unit (14) comprises a traffic sign detector (31) adapted to detect traffic signs in images (30) captured by said imaging apparatus (11) through image processing, a decision section (35) and a traffic sign estimator (36) that is adapted to estimate validity information (37) of one or more traffic signs in an image (30) captured by said imaging apparatus (11).

Abstract: A headlight control system (10) for a motor vehicle including, a controllable headlight (24) adapted to generate variable illumination of the vehicle environment, an imaging apparatus (11) adapted to capture images (100) from a region in front of the motor vehicle, and a data processing device (14) adapted to perform image processing of images (100) captured by the imaging apparatus (11) and to vary the light characteristics of the controllable headlight (24) depending on the image processing. A machine learning model (27) is implemented in the data processing device (14) which is trained to estimate and output an output signal (29) representing a desired illumination of the vehicle environment from one or more images (28) received as input from the imaging apparatus (11).

Abstract: A vision system (1) for a motor vehicle, includes a sensing device (40) adapted to sense data (99), a data processing device (7) adapted to process the sensed data (99), wherein the data processing device (7) includes a controller (30) adapted to switch a power mode of the data processing device (7) between a low-power mode and a high-power mode, wherein a high-power application (110) can be executed in the high-power mode. The data processing device (7), in a low-power mode, is adapted to execute a power mode classifier (20), and the power mode classifier (20) is adapted to classify any input data (99) from the sensing device (40) as requiring executing the high-power application (110) or not, and to output a high-power mode request (130) to the controller (30) to switch the power mode of the data processing device (7) from the low-power mode to the high-power mode in case the power mode classifier (20) has classified the input data (99) as requiring executing the high-power application (110).

Abstract: The present invention relates to the field of digital media recording such as video or audio, where compression is used to reduce the amount of data to save onto a data storage. In particular the invention relates to recording media, wherein the memory area required to store the media is unknown. The invention discloses a method for recording digital streamed media with a number of media frames in a memory by encoding each media frame into an encoded frame comprising a first number of sets of data, representing different characteristics of the media stream. The invention further relates to detecting if the data storage is full and if the data storage is full, storing new frames in the data storage previously occupied by the data sets representing the least prioritized characteristics of the previously stored encoded frames. The invention also relates to a corresponding media recorder and computer program.