Title:
OBJECT DETECTION DEVICE, OBJECT DETECTION SYSTEM, OBJECT DETECTION METHOD, AND NON-TRANSITORY COMPUTER-READABLE MEDIUM IN WHICH PROGRAM IS STORED
Document Type and Number:
WIPO Patent Application WO/2020/217283
Kind Code:
A1
Abstract:
Provided is an object detection device with which it is possible to detect an object with good accuracy. An input image acquisition unit (2) acquires a plurality of input images captured in a plurality of modals differing in a specific detection subject. A perturbed image acquisition unit (4) acquires a plurality of perturbed images in which at least one of the plurality of input images is perturbed. A detection processing unit (6) detects a detection subject included in input images using each of the plurality of perturbed images and the input image among the plurality of input images that is not perturbed, and acquires the detected position of the detection subject and the certainty factor of detection as detection results. An adjustment unit (8) calculates, using an integration parameter, the certainty factor adjusted for each of the perturbed images, on the basis of the detected position and certainty factor acquired for the plurality of perturbed images. An output unit (10) outputs the detection results of the detection subject based on the adjusted certainty factor.
Inventors:
SAWADA AZUSA (JP)
SHIBATA TAKASHI (JP)
SHIBATA TAKASHI (JP)
Application Number:
PCT/JP2019/017072
Publication Date:
October 29, 2020
Filing Date:
April 22, 2019
Export Citation:
Assignee:
NEC CORP (JP)
International Classes:
G06T7/174
Domestic Patent References:
| WO2018147059A1 | 2018-08-16 |
Foreign References:
| JP2017146957A | 2017-08-24 |
Other References:
JINGJING LIU, SHAOTING ZHANG, SHU WANG, DIMITRIS N. METAXAS: "Multispectral Deep Neural Networks for Pedestrian Detection", PROCEEDINGS OF THE BRITISH MACHINE VISION CONFERENCE, September 2016 (2016-09-01), pages 1 - 13, XP055758190, Retrieved from the Internet [retrieved on 20190618]
Attorney, Agent or Firm:
IEIRI Takeshi (JP)
Download PDF: