公开(公告)号：US11776292B2

公开(公告)日：2023-10-03

申请号：US17249803

申请日：2021-03-15

Applicant: Wistron Corp.

Inventor： Kuo-Lun Huang

IPC: G06V40/10 ,  G06T7/246 ,  G06N3/08 ,  G06F18/214 ,  G06N3/045

CPC classification number: G06V40/10 ,  G06F18/214 ,  G06N3/045 ,  G06N3/08 ,  G06T7/248 ,  G06T2207/20081 ,  G06T2207/20084

Abstract: An object identification method includes: generating a tracking sample and an adversarial sample; training a teacher model according to the tracking sample; and initializing a student model according to the teacher model. The student model adjusts a plurality of parameters according to the teacher model and the adversarial sample, in response to the vector difference between the output result of the student model and the output result of the teacher model being lower than the learning threshold, the student model is deemed to have completed training, and the student model is extracted as an object identification model.

公开(公告)号：US20220198181A1

公开(公告)日：2022-06-23

申请号：US17249803

申请日：2021-03-15

Applicant: Wistron Corp.

Inventor： Kuo-Lun Huang

IPC: G06K9/00 ,  G06K9/62 ,  G06T7/246 ,  G06N3/08 ,  G06N3/04

Abstract: An object identification method includes: generating a tracking sample and an adversarial sample; training a teacher model according to the tracking sample; and initializing a student model according to the teacher model. The student model adjusts a plurality of parameters according to the teacher model and the adversarial sample, in response to the vector difference between the output result of the student model and the output result of the teacher model being lower than the learning threshold, the student model is deemed to have completed training, and the student model is extracted as an object identification model.