Facial recognition technology for improving driver safety
Methods for performing operations for improving driver safety across a fleet of vehicles are disclosed. A plurality of safety events pertaining to a driving of a fleet of vehicles by a plurality of drivers are detected. A subset of the events is identified. The subset corresponds to one or more safety events of the plurality of safety events involving one or more vehicles of the fleet of vehicles to which drivers have not been assigned. A user interface is generated for presentation on a client device, the user interface including an interactive user interface element for accessing the subset of the events. One or more user interface elements are provided for accepting or rejecting recommendations for assignments of one of the plurality of drivers to each of the vehicles. The recommendations are generated based on an application of a machine-learned model to images of faces captured.
1 . A system comprising:
one or more computer processors;
one or more computer memories; and
a set of instructions incorporated into the one or more computer processors, the set of instructions configuring the one or more computer processors to perform operations, the operations comprising:
storing a plurality of driver records in a database, each driver record associated with a different driver of a fleet of vehicles and indicating hours of service driven by the corresponding driver;
receiving an image of a face captured during a safety event inside a vehicle from the fleet of vehicles;
predicting an identity of a driver based on the image using a machine-learned facial recognition model;
receiving a human confirmation or correction of the identity of the driver predicted using the machine-learned facial recognition model;
in response to receiving the human confirmation, updating a driver record of the plurality of driver records in the database that corresponds to the predicted identity of the driver, the driver record updated to include one or more hours of service to maintain regulatory compliance; and
updating training of the machine-learned facial recognition model based on the human confirmation or correction of the predicted identity of the driver.
2 . The system of claim 1 , wherein predicting of the identity of the driver comprises processing the image to detect facial features of the face.
3 . The system of claim 2 , wherein the predicting of the identity of the driver comprises comparing the detected facial features to stored facial features associated with known drivers.
4 . The system of claim 3 , wherein the predicting of the identity of the driver comprises determining a correspondence between the detected facial features and the stored facial features that transgresses a threshold level of correspondence.
5 . The system of claim 1 , wherein the human confirmation or correction is from an administrator.
6 . The system of claim 5 , wherein the updating of the training comprises using the received human confirmation or correction as training data to improve future matches between faces capture in images and known drivers.
7 . The system of claim 6 , wherein predicting of the identity of the driver comprises:
receiving a series of images capturing the face of the driver from different viewing angles during operation of the vehicle;
adding one or more of the series of images to a club of images associated with the driver, wherein each image in the club has a threshold degree of similarity to other images in the club; and
using the club of images to train the machine-learned facial recognition model.
8 . A method comprising:
storing a plurality of driver records in a database, each driver record associated with a different driver of a fleet of vehicles and indicating hours of service driven by the corresponding driver;
receiving an image of a face captured during a safety event inside a vehicle from the fleet of vehicles;
predicting an identity of a driver based on the image using a machine-learned facial recognition model;
receiving a human confirmation or correction of the identity of the driver predicted using the machine-learned facial recognition model;
in response to receiving the human confirmation, updating a driver record of the plurality of driver records in the database that corresponds to the predicted identity of the driver, the driver record updated to include one or more hours of service to maintain regulatory compliance; and
updating training of the machine-learned facial recognition model based on the human confirmation or correction of the predicted identity of the driver.
9 . The method of claim 8 , wherein predicting of the identity of the driver comprises processing the image to detect facial features of the face.
10 . The method of claim 9 , wherein the predicting of the identity of the driver comprises comparing the detected facial features to stored facial features associated with known drivers.
11 . The method of claim 10 , wherein the predicting of the identity of the driver comprises determining a correspondence between the detected facial features and the stored facial features that transgresses a threshold level of correspondence.
12 . The method of claim 8 , wherein the human confirmation or correction is from an administrator.
13 . The method of claim 12 , wherein the updating of the training comprises using the received human confirmation or correction as training data to improve future matches between faces capture in images and known drivers.
14 . The method of claim 8 , wherein predicting of the identity of the driver comprises:
receiving a series of images capturing the face of the driver from different viewing angles during operation of the vehicle;
adding one or more of the series of images to a club of images associated with the driver, wherein each image in the club has a threshold degree of similarity to other images in the club; and
using the club of images to train the machine-learned facial recognition model.
15 . A non-transitory machine-readable storage medium comprising instructions that, when executed by one or more processors of a machine, cause the machine to perform operations, the operations comprising:
storing a plurality of driver records in a database, each driver record associated with a different driver of a fleet of vehicles and indicating hours of service driven by the corresponding driver;
receiving an image of a face captured during a safety event inside a vehicle from the fleet of vehicles;
predicting an identity of a driver based on the image using a machine-learned facial recognition model;
receiving a human confirmation or correction of the identity of the driver predicted using the machine-learned facial recognition model;
in response to receiving the human confirmation, updating a driver record of the plurality of driver records in the database that corresponds to the predicted identity of the driver, the driver record updated to include one or more hours of service to maintain regulatory compliance; and
updating training of the machine-learned facial recognition model based on the human confirmation or correction of the predicted identity of the driver.
16 . The non-transitory machine-readable storage medium of claim 15 , wherein predicting of the identity of the driver comprises processing the image to detect facial features of the face.
17 . The non-transitory machine-readable storage medium of claim 16 , wherein the predicting of the identity of the driver comprises comparing the detected facial features to stored facial features associated with known drivers.
18 . The non-transitory machine-readable storage medium of claim 17 , wherein the predicting of the identity of the driver comprises determining a correspondence between the detected facial features and the stored facial features that transgresses a threshold level of correspondence.
19 . The non-transitory machine-readable storage medium of claim 15 , wherein the human confirmation or correction is from an administrator.
20 . The non-transitory machine-readable storage medium of claim 19 , wherein the updating of the training comprises using the received human confirmation or correction as training data to improve future matches between faces capture in images and known drivers.