Paper Abstract and Keywords |
Presentation |
2019-09-20 11:40
Accurate Pedestrian Detection in Thermal Images for FPGA Ryosuke Kuramochi, Masayuki Shimoda, Youki Sada, Shimpei Sato, Hiroki Nakahara (titech) RECONF2019-26 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
Since thermal cameras can detect the heat of objects, they can be used even if there is no light.
Therefore, object detection with a thermal camera is reliable and attracts much attention.
However, it is hard to obtain high detection accuracy with thermal images due to various drawbacks of them.
And also, when incorporating object detection into embedded systems such as a surveillance camera, there is a trade-off between accuracy and processing time.
To realize the real-time pedestrian detection system with high accuracy on FPGA, we propose to use a pair of a thermal image, and a background-subtracted one as an input image for YOLOv2 object detector.
We also propose a pedestrian detector system which contains a vectorizing unit.
With the vectorizing unit, we can store only valid values of a feature map together in the same memory address, resulting in calculations with a high degree of parallelism.
Compared to the conventional pedestrian detection system for thermal images, our method obtains 29 points higher F-score and operates 3.3 times faster.
From these results, our pedestrian detection system is more suitable for embedded surveillance systems. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
Deep Learning / CNN / FPGA / Thermal Image / / / / |
Reference Info. |
IEICE Tech. Rep., vol. 119, no. 208, RECONF2019-26, pp. 31-36, Sept. 2019. |
Paper # |
RECONF2019-26 |
Date of Issue |
2019-09-12 (RECONF) |
ISSN |
Online edition: ISSN 2432-6380 |
Copyright and reproduction |
All rights are reserved and no part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Notwithstanding, instructors are permitted to photocopy isolated articles for noncommercial classroom use without fee. (License No.: 10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
Download PDF |
RECONF2019-26 |
Conference Information |
Committee |
RECONF |
Conference Date |
2019-09-19 - 2019-09-20 |
Place (in Japanese) |
(See Japanese page) |
Place (in English) |
KITAKYUSHU Convention Center |
Topics (in Japanese) |
(See Japanese page) |
Topics (in English) |
Reconfigurable Systems, etc. |
Paper Information |
Registration To |
RECONF |
Conference Code |
2019-09-RECONF |
Language |
Japanese |
Title (in Japanese) |
(See Japanese page) |
Sub Title (in Japanese) |
(See Japanese page) |
Title (in English) |
Accurate Pedestrian Detection in Thermal Images for FPGA |
Sub Title (in English) |
|
Keyword(1) |
Deep Learning |
Keyword(2) |
CNN |
Keyword(3) |
FPGA |
Keyword(4) |
Thermal Image |
Keyword(5) |
|
Keyword(6) |
|
Keyword(7) |
|
Keyword(8) |
|
1st Author's Name |
Ryosuke Kuramochi |
1st Author's Affiliation |
Tokyo Institute of Technology (titech) |
2nd Author's Name |
Masayuki Shimoda |
2nd Author's Affiliation |
Tokyo Institute of Technology (titech) |
3rd Author's Name |
Youki Sada |
3rd Author's Affiliation |
Tokyo Institute of Technology (titech) |
4th Author's Name |
Shimpei Sato |
4th Author's Affiliation |
Tokyo Institute of Technology (titech) |
5th Author's Name |
Hiroki Nakahara |
5th Author's Affiliation |
Tokyo Institute of Technology (titech) |
6th Author's Name |
|
6th Author's Affiliation |
() |
7th Author's Name |
|
7th Author's Affiliation |
() |
8th Author's Name |
|
8th Author's Affiliation |
() |
9th Author's Name |
|
9th Author's Affiliation |
() |
10th Author's Name |
|
10th Author's Affiliation |
() |
11th Author's Name |
|
11th Author's Affiliation |
() |
12th Author's Name |
|
12th Author's Affiliation |
() |
13th Author's Name |
|
13th Author's Affiliation |
() |
14th Author's Name |
|
14th Author's Affiliation |
() |
15th Author's Name |
|
15th Author's Affiliation |
() |
16th Author's Name |
|
16th Author's Affiliation |
() |
17th Author's Name |
|
17th Author's Affiliation |
() |
18th Author's Name |
|
18th Author's Affiliation |
() |
19th Author's Name |
|
19th Author's Affiliation |
() |
20th Author's Name |
|
20th Author's Affiliation |
() |
Speaker |
Author-1 |
Date Time |
2019-09-20 11:40:00 |
Presentation Time |
20 minutes |
Registration for |
RECONF |
Paper # |
RECONF2019-26 |
Volume (vol) |
vol.119 |
Number (no) |
no.208 |
Page |
pp.31-36 |
#Pages |
6 |
Date of Issue |
2019-09-12 (RECONF) |
|