Abstract: A system comprising one or more processors and one or more non-transitory computer-readable media storing computing instructions that, when executed on the one or more processors, cause the one or more processors to perform operations including: receiving sensor data detected by one or more image sensors in a vehicle, wherein the sensor data is representative of driver movements of a driver in the vehicle; categorizing the sensor data as driver postures representative of actions of the driver in the vehicle; rotating and scaling the driver postures to be standardized for different drivers and for different locations of the one or more image sensors within different vehicles; analyzing the sensor data to determine a reference position of the driver in the vehicle; and storing, in a database, the driver postures, as rotated and scaled, and the reference position of the driver in the vehicle. Other embodiments are disclosed.

Abstract: A gate-based vehicle image capture system for analyzing vehicle image data is presented. The system may include a portable imaging gate apparatus configured to capture vehicle image data of a vehicle. The portable imaging gate apparatus may include a plurality of imaging assemblies positioned at a plurality of viewing angles. The system may further include an external processing server configured to receive the vehicle image data from the portable imaging gate apparatus. The external processing server may also analyze the vehicle image data to identify a plurality of vehicle features, and determine a first vehicle feature from the plurality of vehicle features. The first vehicle feature may be related to a vehicle incident. The system may also include a provider server configured to receive the first vehicle feature from the external processing server, and update an aspect of a risk evaluation based on the first vehicle feature.

Abstract: Systems and methods for dynamically assessing vehicle operation are described. According to certain aspects, an electronic device may accumulate telematics data associated with operation of a vehicle, where the telematics data indicates a location of the vehicle. The electronic device may retrieve additional telematics data of additional vehicles operating at or near the location, and compare the accumulated telematics data to the additional telematics data. The electronic device may dynamically generate and display a data visualization indicating the analysis for review by an operator of the vehicle.

Abstract: Apparatuses, systems and methods are provided for generating and transmitting data representative of a vehicle operation mode. More particularly, apparatuses, systems and methods are provided for generating data representative of a vehicle operation mode based on vehicle interior image data.

Abstract: Apparatuses, systems and methods are provided for determining vehicle driver distractions. More particularly, apparatuses, systems and methods are provided for determining distracted drivers associated with vehicle driving routes based on postures of vehicle occupants.

Abstract: Method and system for detecting vehicle occupant actions are disclosed. For example, the method includes receiving, by one or more processors, previously classified image data from a previously classified image database, the previously classified image data representing driver postures that are rotated and scaled to be standardized for a range of different drivers and different locations of a vehicle interior sensor within a given vehicle, wherein the previously classified image data is representative of previously classified vehicle occupant actions; receiving, by the one or more processors, current image data from the vehicle interior sensor subsequent to the vehicle interior sensor being registered within the given vehicle, wherein the current image data is representative of current vehicle occupant actions; and determining, by the one or more processors, a vehicle occupant action based at least in part upon a comparison of the current image data and the previously classified image data.

Abstract: Systems and methods detecting onboard sensor tampering are disclosed. According to embodiments, data captured by interior sensors within a vehicle may be analyzed to determine an indication that the activity of the vehicle operator either cannot be sufficiently detected or cannot be sufficiently identified using the captured data (e.g., that the captured data may be compromised). A date and time associated with the indication may be recorded, and a vehicle operator associated with the indication may be identified. A possible cause for the compromised data may be diagnosed, and notification may be generated indicating that the activity of the vehicle operator either cannot be sufficiently detected or cannot be sufficiently identified, and/or the possible cause. Additionally, a recommendation for restoring sensor functionality may be generated for the vehicle operator based the possible cause.

Abstract: Provided herein is a computer system for creating driving challenges for drivers. The computer system may include a processor in communication with a memory device, and the processor may be programmed to: (i) receive driving data associated with a driver, (ii) generate a first model that models the driving data associated with the driver, (iii) calculate a predicted driving score for the driver based at least in part upon the first model, (iv) generate a second model that predicts a confidence of the predicted driving score, (v) calculate a confidence value of the predicted driving score, wherein the confidence value is a squared error of the predicted driving score, and (vi) generate at least one driving challenge for the driver based at least in part upon the predicted driving score and the confidence value for that predicted driving score.

Abstract: A method of training and using a machine learning model that controls for consideration of undesired factors which might otherwise be considered by the trained model during its subsequent analyses of new data. For example, the model may be a neural network trained on a set of training images to evaluate an insurance applicant based upon an image or audio data of the insurance applicant as part of an underwriting process to determine an appropriate life or health insurance premium. The model is trained to probabilistically correlate an aspect of the applicant's appearance with a personal and/or health-related characteristic. Any undesired factors, such as age, sex, ethnicity, and/or race, are identified for exclusion. The trained model receives the image (e.g., a “selfie”) of the insurance applicant, analyzes the image without considering the identified undesired factors, and suggests the appropriate insurance premium based only on the remaining desired factors.

Abstract: Technology for identifying potentially fraudulent user behavior related to the capture and transmission of vehicle trip telemetry data is provided. Automatic and periodic attempts may be made to wake up (e.g., launch, or change to an active mode) an application installed on a mobile device configured to transmit mobile device data, including captured sensor data, to a remote computer device. Based on the sensor data, the remote computer device may determine a number, frequency, and/or duration of vehicle trips associated with the user of the mobile device. When one or more types of expected mobile device data are not received by the remote computer device after a particular attempt to wake the application, data indicating a failure to receive expected mobile device data may be recorded in a log. Based on the log, an alert indicating potentially fraudulent behavior associated with the user of the mobile device may be generated.

Abstract: A computer-implemented method for improving compression of predictive models includes generating an unlabeled simulated data set by expanding an initial data set, and generating a labeled data set by predicting the unlabeled, simulated data set using a complex model to output a plurality of labels. The method also includes training a relatively simple neural network using the labeled data set.

Abstract: Technology for identifying potentially fraudulent user behavior related to the capture and transmission of vehicle trip telemetry data is provided. Automatic and periodic attempts may be made to wake up (e.g., launch, or change to an active mode) an application installed on a mobile device configured to transmit mobile device data, including captured sensor data, to a remote computer device. Based on the sensor data, the remote computer device may determine a number, frequency, and/or duration of vehicle trips associated with the user of the mobile device. When one or more types of expected mobile device data are not received by the remote computer device after a particular attempt to wake the application, data indicating a failure to receive expected mobile device data may be recorded in a log. Based on the log, an alert indicating potentially fraudulent behavior associated with the user of the mobile device may be generated.

Abstract: A video analytics based image verification system for obtaining initial vehicle profiles is presented. The system may include an external processing server that receives a location of a vehicle and proximate traffic information to determine whether it is safe for a user to obtain an initial vehicle profile. The external processing server may further determine first and second profile features from video data indicative of the vehicle. The external processing server may compare the second profile feature to an image verification indicator to generate an image verification score. A provider server may receive the first profile feature and the image verification score from the external processing server, and update a risk evaluation to include the initial vehicle profile if the image verification score is above an image verification threshold.

Abstract: Systems and methods for improving vehicular safety are provided. According to certain aspects, an electronic device may receive and analyze image data depicting an individual located within a vehicle. The electronic device may also access certain data related to a condition of the individual, and may accordingly determine whether the individual is fit to operate the vehicle. If the individual is unfit to operate the vehicle, the electronic device may alter autonomous operation of the vehicle.

Abstract: Provided herein is a computer system for creating driving challenges for drivers. The computer system may include a processor in communication with a memory device, and the processor may be programmed to: (i) receive driving data associated with a driver, (ii) generate a first model that models the driving data associated with the driver, (iii) calculate a predicted driving score for the driver based at least in part upon the first model, (iv) generate a second model that predicts a confidence of the predicted driving score, (v) calculate a confidence value of the predicted driving score, wherein the confidence value is a squared error of the predicted driving score, and (vi) generate at least one driving challenge for the driver based at least in part upon the predicted driving score and the confidence value for that predicted driving score.

Abstract: A method of training and using a machine learning model that controls for consideration of undesired factors which might otherwise be considered by the trained model during its subsequent analyses of new data. For example, the model may be a neural network trained on a set of training images to evaluate an insurance applicant based upon an image or audio data of the insurance applicant as part of an underwriting process to determine an appropriate life or health insurance premium. The model is trained to probabilistically correlate an aspect of the applicant's appearance with a personal and/or health-related characteristic. Any undesired factors, such as age, sex, ethnicity, and/or race, are identified for exclusion. The trained model receives the image (e.g., a “selfie”) of the insurance applicant, analyzes the image without considering the identified undesired factors, and suggests the appropriate insurance premium based only on the remaining desired factors.

Abstract: A computer-implemented method for improving compression of predictive models includes generating an unlabeled simulated data set by expanding an initial data set, and generating a labeled data set by predicting the unlabeled, simulated data set using a complex model to output a plurality of labels. The method also includes training a relatively simple neural network using the labeled data set.

Abstract: Systems and methods detecting onboard sensor tampering are disclosed. According to embodiments, data captured by interior sensors within a vehicle may be analyzed to determine an indication that the activity of the vehicle operator either cannot be sufficiently detected or cannot be sufficiently identified using the captured data (e.g., that the captured data may be compromised). A date and time associated with the indication may be recorded, and a vehicle operator associated with the indication may be identified. A possible cause for the compromised data may be diagnosed, and notification may be generated indicating that the activity of the vehicle operator either cannot be sufficiently detected or cannot be sufficiently identified, and/or the possible cause. Additionally, a recommendation for restoring sensor functionality may be generated for the vehicle operator based the possible cause.

Abstract: The present disclosure is directed to apparatuses, systems and methods for automatically classifying images of occupants inside a vehicle. More particularly, the present disclosure is directed to apparatuses, systems and methods for automatically classifying images of occupants inside a vehicle by comparing current image feature data to previously classified image features.

Abstract: A system and methods relate to, inter alia, aggregating geolocation data and auxiliary data associated with a plurality of driving activities. The system and methods may further identify the plurality of driving activities that have common geolocation data points based on a geolocation threshold. The system and methods may further determine, from among the identified plurality of driving activities, one or more designated driving activities based on an auxiliary threshold. The system and methods may further generate the neighborhood map that includes the one or more designated driving activities. The system and methods may further transmit a visual representation of, or electronic message detailing, the neighborhood map to a mobile device of a user for display, or otherwise causing the neighborhood map to be displayed to the user.