Presentation 2015-09-18
Comparison of machine learning classifiers for HOG-based human detection on an FPGA
Masahito Oishi, Yoshiki Hayashida, Ryo Fujita, Yuichiro Shibata, Kiyoshi Oguri,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In this paper, we compare Real AdaBoost and a linear SVM from a view point of FPGA implementation of an image-based human detection system. Focusing on the detection rate, the Real AdaBoost was slightly better than the linear SVM. However, linear SVM showed comparable performance with the Real AdaBoost by increasing training data. In terms of FPGA implementation, the number of SVM vectors does not depend on learning data size, while the number of classifiers of Real AdaBoost increases with data size. Due to the linear SVM requires a lot of long weight vectors, smaller implementation was possible by the Real AdaBoost with a straightforward approach shown in this paper.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) FPGA / HOG feature / human detection / AdaBoost / SVM
Paper # RECONF2015-34
Date of Issue 2015-09-11 (RECONF)

Conference Information
Committee RECONF
Conference Date 2015/9/18(2days)
Place (in Japanese) (See Japanese page)
Place (in English) Ehime University
Topics (in Japanese) (See Japanese page)
Topics (in English) Reconfigurable Systems, etc.
Chair Minoru Watanabe(Shizuoka Univ.)
Vice Chair Masato Motomura(Hokkaido Univ.) / Yuichiro Shibata(Nagasaki Univ.)
Secretary Masato Motomura(Toshiba) / Yuichiro Shibata(Univ. of Tsukuba)
Assistant Kazuya Tanikagawa(Hiroshima City Univ.) / Takefumi Miyoshi(e-trees.Japan)

Paper Information
Registration To Technical Committee on Reconfigurable Systems
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Comparison of machine learning classifiers for HOG-based human detection on an FPGA
Sub Title (in English)
Keyword(1) FPGA
Keyword(2) HOG feature
Keyword(3) human detection
Keyword(4) AdaBoost
Keyword(5) SVM
1st Author's Name Masahito Oishi
1st Author's Affiliation Nagasaki University(Nagasaki Univ.)
2nd Author's Name Yoshiki Hayashida
2nd Author's Affiliation Nagasaki University(Nagasaki Univ.)
3rd Author's Name Ryo Fujita
3rd Author's Affiliation Nagasaki University(Nagasaki Univ.)
4th Author's Name Yuichiro Shibata
4th Author's Affiliation Nagasaki University(Nagasaki Univ.)
5th Author's Name Kiyoshi Oguri
5th Author's Affiliation Nagasaki University(Nagasaki Univ.)
Date 2015-09-18
Paper # RECONF2015-34
Volume (vol) vol.115
Number (no) RECONF-228
Page pp.pp.13-18(RECONF),
#Pages 6
Date of Issue 2015-09-11 (RECONF)