公开(公告)号：US20200160106A1

公开(公告)日：2020-05-21

申请号：US16193970

申请日：2018-11-16

Applicant: QUALCOMM Incorporated

Inventor： Ravishankar SIVALINGAM ,  Edwin Chongwoo PARK

IPC: G06K9/62 ,  G06T5/00 ,  G06T7/70

Abstract: Cameras with large field-of-view lenses can cause significant geometric distortions of the images acquired. Training for object detection normally takes place on undistorted images. Thus, in order to detect objects of interest within the acquired images, an undistortion procedure is applied on the acquired images and an object detection is then performed on the undistorted images. Unfortunately, such undistortion procedure is too computationally expensive to be run on some edge devices. To remove the need to perform the undistortion procedure, it is proposed to train for object detection directly from distorted acquired images.

公开(公告)号：US20180336698A1

公开(公告)日：2018-11-22

申请号：US15980447

申请日：2018-05-15

Applicant: QUALCOMM Incorporated

Inventor： Edwin Chongwoo PARK ,  Ravishankar SIVALINGAM

IPC: G06T7/73 ,  G01C3/08 ,  G06N3/08

CPC classification number: G06T7/73 ,  G01C3/08 ,  G06K9/2027 ,  G06K9/2063 ,  G06N3/08 ,  G06N20/00

Abstract: Techniques are presented for detecting a visual marker. A first image containing the visual marker may be captured at a first time under a first lighting condition. A first image-based detection for the visual marker may be performed based on the first image, using a first detector, to produce a first set of results. A second image containing the visual marker may be captured at a second time under a second lighting condition different from the first lighting condition. Based on the first set of results, a second image-based detection for the visual marker may be performed based on the second image, using a second detector different from the first detector, to produce a second set of results.