Data selection based on uncertainty quantification
AssigneeNVIDIA Corporation
InventorsJonathan Tremblay, Fabio Tozeto Ramos, Yuke Zhu, Anima Anandkumar, Guanya Shi
Life of this patent
Estimating the term…
Abstract
Apparatuses, systems, and techniques generate poses of an object based on data of the object observed from a first viewpoint and a second viewpoint. The poses can be evaluated to determine a portion of the data usable by an estimator to generate a pose of the object.
Claims31
A computer system comprising one or more processors and computer-readable memory storing instructions executable by the one or more processors to cause the computer system to at least:
- perform a first comparison of first results of a first estimator and a second estimator, the first results obtained by inputting into the first estimator and second estimator first data associated with observing an object;
- perform a second comparison of second results of the first estimator and the second estimator, the second results obtained by inputting into the first estimator and the second estimator second data associated with observing the object;
- and
- based at least in part on the first comparison and the second comparison, select the first data associated with observing the object or the second data associated with observing the object to be obtained by the first estimator or the second estimator.
A device comprising:
- one or more processors and memory storing executable instructions that, as a result of being executed by the one or more processors, cause the device to:
- generate a first set of data based on an object observed from a first viewpoint;
- generate a second set of data based on the object observed from a second viewpoint;
- determine a first value based on the first set of data and a second value based on the second set of data;
- and
- based on the first and second values, selecting the first set of data or the second set of data to be processed.
A computer-implemented method comprising:
- generating a plurality of evaluations by at least, for each viewpoint of a plurality of viewpoints, evaluating outputs of estimators applied to data associated with an object detected from the viewpoint;
- and
- determining to process, based at least in part on the plurality of evaluations, the data associated with the object detected from a viewpoint of a plurality of viewpoints.
A non-transitory machine-readable medium having stored thereon a set of instructions, which if performed by one or more processors, cause the one or more processors to at least:
- obtain a first set of object poses generated by a plurality of estimators based on data;
- obtain a second set of object poses generated by the plurality of estimators based on the data;
- evaluate the object poses of the first and second sets of poses to select which of the first set of poses or the second set of poses comprises poses most similar to one another;
- and
- based on the first set of poses or the second set of poses selected by evaluating the object poses, store in computer-implemented storage at least a portion of the data, the portion of the data usable to generate an object pose associated with an object.
Dependencies are inferred from the claim text.