US 12,586,365 B2 · Object classification using multiple labels for autonomous systems and applications

Abstract

In various examples, multilabel hierarchical classification of objects for autonomous systems and applications is described herein. Systems and methods are disclosed that use one or more neural networks to classify objects, such as traffic signs, using multilabel classification and/or hierarchical classification. For instance, a multilabel subnetwork of the neural network(s) may classify an object based at least on one or more attributes associated with the object. As such, the output from the multilabel subnetwork may include at least a classification associated with the object and an attribute classification(s) associated with the object. A hierarchical subnetwork of the neural network(s) may also classify the object using one or more class labels, where a class label indicates another classification and/or a class group associated with the object. The systems and methods may then use the classification, the attribute classification(s), and/or the class label(s) to determine a final classification associated with the object.

Claims20

Dependent claims

A method comprising:

determining, using one or more neural networks and based at least on sensor data corresponding to a sign, data associated with one or more attribute classifications of the sign;
determining, using one or more first layers of the one or more neural networks and based at least on the data, a first output indicating one or more sign classifications associated with the sign;
determining, using one or more second layers of the one or more neural networks that are different from the one or more first layers, and based at least on the data, a second output indicating at least an attribute classification, from the one or more attribute classifications, associated with the sign;
and
causing, based at least on the first output and the second output, a machine to perform one or more operations.

2 more elements

A system comprising:

one or more processors to:
determine, using one or more neural networks and based at least on sensor data representative of an object, a first output indicating one or more object classifications for classifying the object;
determine, using the one or more neural networks and based at least on the sensor data representative of the object, a second output associated with a hierarchy corresponding to the object, the hierarchy indicating at least one or more group classifications along with the one or more object classifications that are included in a group classification of the one or more group classifications;
determine, based at least on the first output and the second output, that an object classification from the one or more object classifications is associated with the object;
and
cause, based at least on the object classification, a machine to perform one or more operations.

3 more elements

One or more processors comprising processing circuitry to:

determine, using one or more neural networks and based at least on sensor data corresponding to an object, a first output indicating one or more class groups associated with the object and a second output indicating one or more object classifications associated with the object;
determine, based at least on the first output, a class group from the one or more class groups;
determine, based at least on the second output, an object classification from the one or more object classifications that is included in the class group;
and
determine, based at least on the object classification being included in the class group, that the object classification includes a final object classification associated with the object.

2 more elements

Dependencies are inferred from the claim text.

Object classification using multiple labels for autonomous systems and applications

Life of this patent

Object classification using multiple labels for autonomous systems and applications

Life of this patent

Abstract

Claims20