Delegation of object and pose detection

1. A method implemented using one or more processors, comprising:

operating an object recognition client to facilitate object recognition for a robot;

receiving, by the object recognition client, sensor data indicative of an observed object in an environment;

generating, by the object recognition client, a soft classifier of a type of the observed object, wherein the soft classifier is generated based on both the sensor data and a type of location in which the object is observed;

providing, by the object recognition client, to a root object recognition server operating remotely from the robot, data indicative of the observed object, wherein the data indicative of the observed object includes the soft classifier;

receiving, by the object recognition client, from one or more targeted object recognition modules selected by the root object recognition server based on the soft classifier, one or more inferences about an object type or pose of the observed object; and

determining, by the object recognition client, information about the observed object based on the one or more inferences.

2. The method of claim 1 , wherein the type of location in which the object is observed is obtained from an offline resource that is available to the robot while the robot is not in network communication with the root object recognition server.

3. The method of claim 1 , wherein the soft classifier is generated further based on a location of the robot.

4. The method of claim 1 , wherein the soft classifier is generated further based on a time of day at the robot's location.

5. The method of claim 1 , wherein the soft classifier is generated further based on an attribute of a task being performed by the robot.

6. The method of claim 1 , wherein the one or more processors are integral with the robot.

7. The method of claim 1 , wherein the one or more processors are communicatively coupled with the robot via a local area or personal area network.

8. A robot comprising one or more processors and memory storing instructions that, in response to execution of the instructions by the one or more processors, cause the one or more processors to:

operate an object recognition client to facilitate object recognition for the robot;

receive, by the object recognition client, sensor data indicative of an observed object in an environment;

generate, by the object recognition client, a soft classifier of a type of the observed object, wherein the soft classifier is generated based on the sensor data and a type of location in which the object is observed;

provide, by the object recognition client, to a root object recognition server operating remotely from the robot, data indicative of the observed object, wherein the data indicative of the observed object includes the soft classifier;

receive, by the object recognition client, from one or more targeted object recognition modules selected by the root object recognition server based on the soft classifier, one or more inferences about an object type or pose of the observed object; and

determine, by the object recognition client, information about the observed object based on the one or more inferences.

9. The robot of claim 8 , wherein the type of location in which the object is observed obtained from an offline resource that is available to the robot while the robot is not in network communication with the root object recognition server.

10. The robot of claim 8 , wherein the soft classifier is generated further based on a location of the robot.

11. The robot of claim 8 , wherein the soft classifier is generated further based on a time of day at the robot's location.

12. The robot of claim 8 , wherein the soft classifier is generated further based on an attribute of a task being performed by the robot.

13. The robot of claim 8 , wherein the one or more processors are integral with the robot.

14. The robot of claim 8 , wherein the one or more processors are communicatively coupled with the robot via a local area or personal area network.

15. At least one non-transitory computer-readable medium comprising instructions that, in response to execution of the instructions by one or more processors, cause the one or more processors to perform the following operations:

operating an object recognition client to facilitate object recognition for a robot;

receiving, by the object recognition client, sensor data indicative of an observed object in an environment;

generating, by the object recognition client, a soft classifier of a type of the observed object, wherein the soft classifier is generated based on the sensor data and a type of location in which the object is observed;

providing, by the object recognition client, to a root object recognition server operating remotely from the robot, data indicative of the observed object, wherein the data indicative of the observed object includes the soft classifier;

receiving, by the object recognition client, from one or more targeted object recognition modules selected by the root object recognition server based on the soft classifier, one or more inferences about an object type or pose of the observed object; and

determining, by the object recognition client, information about the observed object based on the one or more inferences.

16. The at least one non-transitory computer-readable medium of claim 15 , wherein the type of location in which the object is observed is obtained from an offline resource that is available to the robot while the robot is not in network communication with the root object recognition server.

17. The at least one non-transitory computer-readable medium of claim 15 , wherein the soft classifier is generated further based on a location of the robot.