The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
모바일 로봇에 대한 최근 연구에 따르면 CNN(Convolutional Neural Network)은 특히 대규모 동적 환경의 시각적 장소 인식에서 인상적인 성능을 달성한 것으로 나타났습니다. 그러나 CNN은 로봇 내비게이션에 대한 실시간 요구를 충족할 수 없는 넓은 이미지 표현 공간을 제공합니다. 이 문제를 해결하기 위해 우리는 분산을 통해 CNN 계층에서 얻은 특징 맵의 특징 효율성을 평가하고 핵심 특징 맵을 예약하고 적응형 이진화를 수행하는 새로운 방법을 제안합니다. 실험 결과는 우리 방법의 효과와 효율성을 보여줍니다. 시각적 장소 인식을 위한 최신 방법과 비교할 때 우리의 방법은 정밀도가 크게 손실되지 않을 뿐만 아니라 이미지 표현 공간을 크게 줄입니다.
Yutian CHEN
Army Engineering University of PLA
Wenyan GAN
Army Engineering University of PLA
Shanshan JIAO
Army Engineering University of PLA
Youwei XU
Army Engineering University of PLA
Yuntian FENG
Army Engineering University of PLA
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
부
Yutian CHEN, Wenyan GAN, Shanshan JIAO, Youwei XU, Yuntian FENG, "Salient Feature Selection for CNN-Based Visual Place Recognition" in IEICE TRANSACTIONS on Information,
vol. E101-D, no. 12, pp. 3102-3107, December 2018, doi: 10.1587/transinf.2018EDP7175.
Abstract: Recent researches on mobile robots show that convolutional neural network (CNN) has achieved impressive performance in visual place recognition especially for large-scale dynamic environment. However, CNN leads to the large space of image representation that cannot meet the real-time demand for robot navigation. Aiming at this problem, we evaluate the feature effectiveness of feature maps obtained from the layer of CNN by variance and propose a novel method that reserve salient feature maps and make adaptive binarization for them. Experimental results demonstrate the effectiveness and efficiency of our method. Compared with state of the art methods for visual place recognition, our method not only has no significant loss in precision, but also greatly reduces the space of image representation.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2018EDP7175/_p
부
@ARTICLE{e101-d_12_3102,
author={Yutian CHEN, Wenyan GAN, Shanshan JIAO, Youwei XU, Yuntian FENG, },
journal={IEICE TRANSACTIONS on Information},
title={Salient Feature Selection for CNN-Based Visual Place Recognition},
year={2018},
volume={E101-D},
number={12},
pages={3102-3107},
abstract={Recent researches on mobile robots show that convolutional neural network (CNN) has achieved impressive performance in visual place recognition especially for large-scale dynamic environment. However, CNN leads to the large space of image representation that cannot meet the real-time demand for robot navigation. Aiming at this problem, we evaluate the feature effectiveness of feature maps obtained from the layer of CNN by variance and propose a novel method that reserve salient feature maps and make adaptive binarization for them. Experimental results demonstrate the effectiveness and efficiency of our method. Compared with state of the art methods for visual place recognition, our method not only has no significant loss in precision, but also greatly reduces the space of image representation.},
keywords={},
doi={10.1587/transinf.2018EDP7175},
ISSN={1745-1361},
month={December},}
부
TY - JOUR
TI - Salient Feature Selection for CNN-Based Visual Place Recognition
T2 - IEICE TRANSACTIONS on Information
SP - 3102
EP - 3107
AU - Yutian CHEN
AU - Wenyan GAN
AU - Shanshan JIAO
AU - Youwei XU
AU - Yuntian FENG
PY - 2018
DO - 10.1587/transinf.2018EDP7175
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E101-D
IS - 12
JA - IEICE TRANSACTIONS on Information
Y1 - December 2018
AB - Recent researches on mobile robots show that convolutional neural network (CNN) has achieved impressive performance in visual place recognition especially for large-scale dynamic environment. However, CNN leads to the large space of image representation that cannot meet the real-time demand for robot navigation. Aiming at this problem, we evaluate the feature effectiveness of feature maps obtained from the layer of CNN by variance and propose a novel method that reserve salient feature maps and make adaptive binarization for them. Experimental results demonstrate the effectiveness and efficiency of our method. Compared with state of the art methods for visual place recognition, our method not only has no significant loss in precision, but also greatly reduces the space of image representation.
ER -