SYSTEM AND METHOD FOR AUTOMATIC DRIVER IDENTIFICATION

• US 20180012092A1
• Filed: 07/05/2017
• Published: 01/11/2018
• Est. Priority Date: 07/05/2016
• Status: Active Grant

• Current Assignee: Nauto, Inc.
• Original Assignee: Nauto, Inc.
• Inventors: Gleeson-May, Michael, Soo, Frederick, Heck, Stefan
• Granted Patent: US 10,133,942 B2
• Time in Patent Office: Days
• CPC Class Codes: B60R 25/25 using biometry
• CPC Class Codes: G06Q 40/08 Insurance
• CPC Class Codes: G06V 10/50 by performing operations within image blocks; by using histograms, e.g. histogram of oriented gradients [HoG]; by summing image-intensity values; Projection analysis
• CPC Class Codes: G06V 10/95 structured as a network, e.g. client-server architectures
• CPC Class Codes: G06V 20/597 Recognising the driver's state or behaviour, e.g. attention or drowsiness
• CPC Class Codes: G06V 40/168 Feature extraction; Face representation
• CPC Class Codes: G06V 40/171 Local features and components; Facial parts eye characteristics G06V40/18; Occluding parts, e.g. glasses; Geometrical relationships
• CPC Class Codes: G06V 40/172 Classification, e.g. identification

Abstract

A method for driver identification including recording a first image of a vehicle driver; extracting a set of values for a set of facial features of the vehicle driver from the first image; determining a filtering parameter; selecting a cluster of driver identifiers from a set of clusters, based on the filtering parameter; computing a probability that the set of values is associated with each driver identifier of the cluster; determining, at the vehicle sensor system, driving characterization data for the driving session; and in response to the computed probability exceeding a first threshold probability: determining that the new set of values corresponds to one driver identifier within the selected cluster, and associating the driving characterization data with the one driver identifier.

Claims

1. A method for driver identification comprising:

recording, at a vehicle sensor system and at a first time point, a first image of a vehicle driver;

extracting a set of values for a set of facial features of the vehicle driver from the first image;

determining a filtering parameter;

selecting a cluster of driver identifiers from a set of clusters, based on the filtering parameter;

computing a probability that the set of values is associated with each driver identifier by repeating;

recording a new image of the vehicle driver at a time point during a driving session and after the first time point;

extracting a new set of values for the set of facial features from the new image; and

computing the probability that the new set of values is associated with each driver identifier based on the new set of values;

determining, at the vehicle sensor system, driving characterization data for the driving session;

in response to the computed probability exceeding a first threshold probability;

determining that the new set of values corresponds to one driver identifier within the selected cluster; and

associating the driving characterization data with the one driver identifier; and

in response to the computed probability associated with each driver identifier in the cluster falling below a second threshold probability;

determining a new driver identifier;

associating the new driver identifier with the set of values for the set of facial features of the vehicle driver; and

associating the driving characterization data with the new driver identifier.

2. The method of claim 1, wherein determining the filtering parameter comprises determining that the time point falls within a scheduled time duration, and wherein selecting the cluster based on the filtering parameter comprises selecting the cluster corresponding to the scheduled time duration.

3. The method of claim 1, further comprising adding the new driver identifier to the selected cluster of candidate driver identities in response to determining the new driver identifier.

4. The method of claim 1, further comprising clustering a set of driver identifiers into clusters based on one or more managing entities associated with each of the set of driver identifiers, wherein each cluster corresponds to a single managing entity.

5. The method of claim 1, wherein computing the filtering parameter value comprises computing a cluster probability for each of the set of clusters based on the extracted set of values, wherein the cluster probability comprises a likelihood that a candidate driver identity within a cluster corresponds to the driver identity.

6. The method of claim 1, wherein recording the first image comprises recording a sequence of images and selecting the first image from the sequence of images.

7. The method of claim 6, further comprising selecting the first image from the sequence of images based on an ambient light level extracted from each of the sequence of images.

8. The method of claim 6, further comprising determining the driving characterization data from the sequence of images.

9. The method of claim 8, wherein the driving characterization data comprises occupancy data.

10. The method of claim 1, further comprising recording the first image in response to the occurrence of a sampling event.

11. The method of claim 10, wherein the sampling event comprises a time point contemporaneous with the first time point substantially matching a scheduled driving time.

12. A method for driver identification comprising:

detecting a sampling event at a vehicle sensor system within a vehicle driven by a vehicle driver;

determining a filtering parameter in response to sampling event detection;

selecting a cluster of driver identifiers from a set of clusters, based on the filtering parameter value;

in response to detecting the sampling event, computing a probability that the vehicle driver is associated with a driver identifier of the selected cluster, comprising iteratively;

recording an image of the vehicle driver;

extracting a set of facial feature values from the image; and

updating the probability based on the set of facial feature values;

determining, at the vehicle sensor system, driving characterization data during a driving session contemporaneous with detecting the sampling event; and

in response to the computed probability exceeding a first threshold probability, associating the driving characterization data with the driver identifier.

13. The method of claim 12, wherein recording the image of the vehicle driver comprises recording a sequence of images of the vehicle driver, and extracting the set of facial feature values from the image comprises calculating the set of facial feature values from a selected image of the sequence.

14. The method of claim 13, further comprising selecting the selected image from the sequence of images based on a computed projected area of a head region of the vehicle driver within the image exceeding a threshold projected area.

15. The method of claim 13, wherein the sequence of images is recorded at a sampling frequency, and wherein the sampling frequency is dynamically adjusted based on a highest computed probability that the vehicle driver is associated with each driver identifier of the selected cluster.

16. The method of claim 13, further comprising determining the driving characterization data from the sequence of images, and wherein the driving characterization data comprises at least one of occupancy data and driver behavior data.

17. The method of claim 12, wherein the filtering parameter value comprises a hardware identifier associated with a mobile device of the vehicle driver.

18. The method of claim 12, wherein determining the filtering parameter comprises computing a cluster probability for each of the set of clusters, wherein the cluster probability comprises a likelihood that any driver identifier within a cluster is associated with the vehicle driver, and wherein selecting the cluster comprises selecting the cluster from the set of clusters having a highest computed cluster probability.

19. The method of claim 18, further comprising, in response to the computed probability falling below a second threshold probability:

generating a new driver identifier,

associating the vehicle driver with the new driver identifier,

associating the driving characterization data with the new driver identifier, and

adding the new driver identifier to the selected cluster.

20. The method of claim 12, wherein detecting the sampling event comprises detecting, at the vehicle sensor system, a sensor signal pattern substantially matching a predetermined pattern associated with at least one of a vehicle door opening and a vehicle door closing.