公开(公告)号：US20220147743A1

公开(公告)日：2022-05-12

申请号：US17226584

申请日：2021-04-09

Applicant: Nvidia Corporation

Inventor： Donna Roy ,  Suraj Kothawade ,  Elmar Haussmann ,  Jose Manuel Alvarez Lopez ,  Michele Fenzi ,  Christoph Angerer

IPC: G06K9/00 ,  G06K9/62 ,  G06K9/72 ,  G06N3/08

Abstract: Approaches presented herein provide for semantic data matching, as may be useful for selecting data from a large unlabeled dataset to train a neural network. For an object detection use case, such a process can identify images within an unlabeled set even when an object of interest represents a relatively small portion of an image or there are many other objects in the image. A query image can be processed to extract image features or feature maps from only one or more regions of interest in that image, as may correspond to objects of interest. These features are compared with images in an unlabeled dataset, with similarity scores being calculated between the features of the region(s) of interest and individual images in the unlabeled set. One or more highest scored images can be selected as training images showing objects that are semantically similar to the object in the query image.

公开(公告)号：US12272148B2

公开(公告)日：2025-04-08

申请号：US17226584

申请日：2021-04-09

Applicant: Nvidia Corporation

Inventor： Donna Roy ,  Suraj Kothawade ,  Elmar Haussmann ,  Jose Manuel Alvarez Lopez ,  Michele Fenzi ,  Christoph Angerer

IPC: G06V20/56 ,  G06F18/2113 ,  G06F18/214 ,  G06F18/22 ,  G06N3/08 ,  G06V30/262

Abstract: Approaches presented herein provide for semantic data matching, as may be useful for selecting data from a large unlabeled dataset to train a neural network. For an object detection use case, such a process can identify images within an unlabeled set even when an object of interest represents a relatively small portion of an image or there are many other objects in the image. A query image can be processed to extract image features or feature maps from only one or more regions of interest in that image, as may correspond to objects of interest. These features are compared with images in an unlabeled dataset, with similarity scores being calculated between the features of the region(s) of interest and individual images in the unlabeled set. One or more highest scored images can be selected as training images showing objects that are semantically similar to the object in the query image.