Title:
OBJECT DETECTION METHOD, OBJECT DETECTION DEVICE, AND OBJECT DETECTION PROGRAM
Document Type and Number:
Japanese Patent JP2021033580
Kind Code:
A
Abstract:
To provide an object detection method, an object detection device, and an object detection program for easily and accurately performing object detection using machine learning.SOLUTION: An object detection method includes a preprocessing process for performing, by a preprocessing unit 11a, first preprocessing for shrinking a mask, which is an area to be learned, and second preprocessing for extracting the contour of the mask, a first learning process for generating, by a first learning unit 11b, a first determination model by executing first learning using a learning image subjected to the first preprocessing and a second learning process for generating, by a second learning unit 11c, a second determination model by executing second learning using a learning image subjected to the second preprocessing, a determination process for acquiring, by a determination unit 11d, a first determination result by inputting a determination target image into the first determination model and a second determination result by inputting the determination target image into the second determination model, and a classification process for separating, by a separation unit 11e, individual objects in the determination target image based on the first determination result and the second determination result.SELECTED DRAWING: Figure 2
More Like This:
Inventors:
SENDA SHUJI
OKA AYAKO
ARIYOSHI KOHEI
OKA AYAKO
ARIYOSHI KOHEI
Application Number:
JP2019152218A
Publication Date:
March 01, 2021
Filing Date:
August 22, 2019
Export Citation:
Assignee:
YE DIGITAL CO LTD
International Classes:
G06T7/00; G06N20/00
Domestic Patent References:
| JP2019133433A | 2019-08-08 |
Other References:
IOANA MOCAN: "Automatic Detection of Tumor Cells in Microscopic Images of Unstained Blood using Convolutional Neur", 2018 IEEE 14TH INTERNATIONAL CONFERENCE ON INTELLIGENT COMPUTER COMMUNICATION AND PROCESSING (ICCP), JPN6023016787, 2018, pages 319 - 324, XP033433607, ISSN: 0005051439, DOI: 10.1109/ICCP.2018.8516638
久保田 涼介: "全層畳み込みニューラルネットワークを用いた透明物体の輪郭抽出", 電子情報通信学会技術研究報告 ＶＯＬ．１１８ ＮＯ．３５, vol. SIP2018-10 (2018-05), JPN6023016786, 10 May 2018 (2018-05-10), JP, pages 41 - 46, ISSN: 0005051440
久保田 涼介: "全層畳み込みニューラルネットワークを用いた透明物体の輪郭抽出", 電子情報通信学会技術研究報告 ＶＯＬ．１１８ ＮＯ．３５, vol. SIP2018-10 (2018-05), JPN6023016786, 10 May 2018 (2018-05-10), JP, pages 41 - 46, ISSN: 0005051440
Attorney, Agent or Firm:
Sakai International Patent Office