Back to results
US 12,183,050 B1Granted
Inferencing using neural networks
AssigneeNVIDIA Corporation
InventorsTushar Santosh Khinvasara, Kaustubh Shriharsh Purandare
Life of this patent
Estimating the term…
Abstract
Apparatuses, systems, and techniques to use data obtained from an inferred object to determine whether to re-infer the same inferred object. In at least one embodiment, data obtained from inferencing a tracked object among a set of images is used to adjust a set of conditions.
Claims20
A method, comprising:
- obtaining data corresponding to inferencing of one or more objects tracked among a set of images;
- adjusting, based at least in part on the obtained data, a set of conditions for determining when at least one object of the one or more objects is re-inferenced, wherein the set of conditions comprises satisfying a comparison between a threshold and a computed size representing the at least one object;
- and
- re-inferencing the at least one object when the comparison is satisfied.
1 more element
A system, comprising:
- one or more processors to:
- obtain data corresponding to an inferred object detected from one or more images;
- adjust a set of conditions based, at least in part, on the obtained data, wherein the set of conditions comprises satisfying a comparison between a threshold and a computed size representing the inferred object;
- and
- re-infer the object detected from one or more additional images based, at least in part, on satisfying the comparison.
2 more elements
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:
- infer a set of objects obtained from a first image;
- obtain a second image, the second image comprising a subset of the set of objects;
- determine, from data corresponding to each object from the subset, whether to adjust a set of conditions to be satisfied, wherein the set of conditions comprises satisfying a comparison between a threshold value and a computed size representing an object from the subset;
- and
- re-infer, based on satisfying the comparison, the object from the subset to refine at least one classification parameter of the object.
2 more elements
Dependencies are inferred from the claim text.