System for recognizing object based on self-supervised learning of artificial intelligence and method implementing the same
AssigneesHyundai Motor Company, Kia Corporation
InventorJin Ho PARK
Life of this patent
Estimating the term…
Abstract
An artificial intelligence (AI)-based system and a method for recognizing an object based on at least one autonomous driving sensor. The system includes a processor and memory storing instructions that, when executed, perform operations including obtaining point cloud data from the at least one autonomous driving sensor and clustering the point cloud data into a cluster of points, extracting a latent feature from an image obtained from the at least one autonomous driving sensor, based on a result of a self-supervised learning process, and assigning a matching class to the cluster of points, and generating a pseudo GT bounding box for the cluster of points based on the matching class.
Claims20
An artificial intelligence (AI)-based system configured to perform object recognition, the system comprising:
- at least one processor;
- and
- at least one memory storing computer program instructions that, based on being executed by the at least one processor, perform operations of recognizing an object in a surrounding environment based on at least one autonomous driving sensor, the operations comprising:
- obtaining an image from the at least one autonomous driving sensor, and extracting at least one latent feature from the image, based on a result of a first self-supervised learning process;
- for each latent feature among the at least one latent feature, assigning a class to the latent feature based on a result of a second self-supervised learning process;
- obtaining point cloud data from the at least one autonomous driving sensor, and clustering the point cloud data to generate a cluster of points;
- for each point in the cluster of points, determining a corresponding latent feature among the at least one latent feature;
- assigning a matching class to the cluster of points based on a correspondence between classes assigned to the latent features and points in the cluster of points;
- and
- generating a pseudo ground-truth (GT) bounding box for the cluster of points, based on the matching class.
A method for performing artificial intelligence (AI)-based object recognition based on at least one autonomous driving sensor, the method comprising:
- obtaining an image from the at least one autonomous driving sensor, and extracting at least one latent feature from the image, based on a result of a first self-supervised learning process;
- for each latent feature among the at least one latent feature, assigning a class to the latent feature based on a result of a second self-supervised learning process;
- obtaining point cloud data from the at least one autonomous driving sensor, and clustering the point cloud data to generate a cluster of points;
- for each point in the cluster of points, determining a corresponding latent feature among the at least one latent feature;
- assigning a matching class to the cluster of points based on a correspondence between classes assigned to the latent features and points in the cluster of points;
- and
- generating a pseudo ground-truth (GT) bounding box for the cluster of points, based on the matching class.
Dependencies are inferred from the claim text.