公开(公告)号：US11948387B2

公开(公告)日：2024-04-02

申请号：US17170307

申请日：2021-02-08

Applicant: ADOBE INC.

Inventor： Sumit Shekhar ,  Bhanu Prakash Reddy Guda ,  Ashutosh Chaubey ,  Ishan Jindal ,  Avneet Jain

IPC: G06V40/10 ,  G06F18/21 ,  G06F18/211 ,  G06F18/214 ,  G06N20/00 ,  G06V20/20

CPC classification number: G06V40/10 ,  G06F18/211 ,  G06F18/2155 ,  G06F18/2178 ,  G06N20/00 ,  G06V20/20

Abstract: Systems and methods for training an object detection network are described. Embodiments train an object detection network using a labeled training set, wherein each element of the labeled training set includes an image and ground truth labels for object instances in the image, predict annotation data for a candidate set of unlabeled data using the object detection network, select a sample image from the candidate set using a policy network, generate a labeled sample based on the selected sample image and the annotation data, wherein the labeled sample includes labels for a plurality of object instances in the sample image, and perform additional training on the object detection network based at least in part on the labeled sample.

公开(公告)号：US20220253630A1

公开(公告)日：2022-08-11

申请号：US17170307

申请日：2021-02-08

Applicant: ADOBE INC.

Inventor： Sumit Shekhar ,  Bhanu Prakash Reddy Guda ,  Ashutosh Chaubey ,  Ishan Jindal ,  Avneet Jain

IPC: G06K9/00 ,  G06K9/62 ,  G06N20/00

Abstract: Systems and methods for training an object detection network are described. Embodiments train an object detection network using a labeled training set, wherein each element of the labeled training set includes an image and ground truth labels for object instances in the image, predict annotation data for a candidate set of unlabeled data using the object detection network, select a sample image from the candidate set using a policy network, generate a labeled sample based on the selected sample image and the annotation data, wherein the labeled sample includes labels for a plurality of object instances in the sample image, and perform additional training on the object detection network based at least in part on the labeled sample.