Explainability for event alerts in video data

1 . A method comprising:

representing, by a device, spatial characteristics over time of an object in video data as one or more timeseries;

detecting, by the device, an event based on a rate of change of behavioral regimes associated with different portions of the one or more timeseries;

selecting, by the device, contextual data for the event, the contextual data comprising spatial timeseries information and data identifying one or more anomaly thresholds for different types of objects or different activities, wherein the one or more anomaly thresholds vary based on at least one of a demographic characteristic of the object or an activity being performed by the object; and

providing, by the device, an alert for the event to a user interface regarding the event that includes the contextual data, the contextual data enabling comparison of the rate of change of behavioral regimes against the one or more anomaly thresholds for the different types of objects or different activities.

2 . The method as in claim 1 , wherein the spatial characteristics comprise a detected centroid of the object.

3 . The method as in claim 1 , wherein the different types of objects comprise people having different demographics.

4 . The method as in claim 1 , wherein the different types of objects comprise different types of vehicles.

5 . The method as in claim 1 , wherein the alert is based further on the rate of change of transitions between the behavioral regimes of the object and that of one or more other objects in the video data.

6 . The method as in claim 1 , wherein the behavioral regimes are associated with the object performing different actions.

7 . The method as in claim 1 , wherein the object is a person or vehicle.

8 . The method as in claim 1 , wherein the object is a cluster of people or vehicles.

9 . The method as in claim 1 , wherein the device is an edge device in a network.

10 . An apparatus, comprising:

a network interface to communicate with a computer network;

a processor coupled to the network interface and configured to execute one or more processes; and

a memory configured to store a process that is executed by the processor, the process when executed configured to:

represent spatial characteristics over time of an object in video data as one or more timeseries;

detect an event based on a rate of change of behavioral regimes associated with different portions of the one or more timeseries;

select contextual data for the event, the contextual data comprising timeseries information and data identifying one or more anomaly thresholds for different types of objects or different activities, wherein the one or more anomaly thresholds vary based on at least one of a demographic characteristic of the object or an activity being performed by the object; and

provide an alert for the event to a user interface regarding the event that includes the contextual data, the contextual data enabling comparison of the rate of change of behavioral regimes against the one or more anomaly thresholds for the different types of objects or different activities.

11 . The apparatus as in claim 10 , wherein the spatial characteristics comprise a detected centroid of the object.

12 . The apparatus as in claim 10 , wherein the different types of objects comprise people having different demographics.

13 . The apparatus as in claim 10 , wherein the different types of objects comprise different types of vehicles.

14 . The apparatus as in claim 10 , wherein the alert is based further on the rate of change of transitions between the behavioral regimes of the object and that of one or more other objects in the video data.

15 . The apparatus as in claim 10 , wherein the behavioral regimes are associated with the object performing different actions.

16 . The apparatus as in claim 10 , wherein the object is a person or vehicle.

17 . The apparatus as in claim 10 , wherein the object is a cluster of people or vehicles.

18 . A tangible, non-transitory, computer-readable medium storing program instructions that cause a device to execute a process comprising:

representing, by the device, spatial characteristics over time of an object in video data as one or more timeseries;

detecting, by the device, an event based on a rate of change of behavioral regimes associated with different portions of the one or more timeseries;

selecting, by the device, contextual data for the event, the contextual data comprising spatial timeseries information and data identifying one or more anomaly thresholds for different types of objects or different activities, wherein the one or more anomaly thresholds vary based on at least one of a demographic characteristic of the object or an activity being performed by the object; and

providing, by the device, an alert for the event to a user interface regarding the event that includes the contextual data, the contextual data enabling comparison of the rate of change of behavioral regimes against the one or more anomaly thresholds for the different types of objects or different activities.