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
컨볼루션 신경망을 객체 감지 분야에 포함시키는 과정에서 많은 컴퓨터 비전 작업이 좋은 성공을 거두었습니다. 다양한 규모의 대상을 적응시키기 위해 전통적인 객체 감지 방법이 가우시안 이미지 피라미드에서 서로 다른 객체를 감지하기 때문에 딥 피처 피라미드가 널리 사용됩니다. 그러나 앵커와 대상의 특징 분포 간의 불일치로 인해 다양한 규모의 대상에 대한 정확한 탐지는 여전히 어려운 과제입니다. 이론적인 수용장과 유효수용장의 차이를 고려하여 유효수용장을 기준으로 하는 새로운 앵커 생성 방법을 제안한다. 제안된 방법은 PASCAL VOC 데이터세트에서 평가되었으며 좋은 결과를 보였다.
Baojun ZHAO
Beijing Institute of Technology
Boya ZHAO
Beijing Institute of Technology
Linbo TANG
Beijing Institute of Technology
Baoxian WANG
Shijiazhuang Tiedao University
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
Baojun ZHAO, Boya ZHAO, Linbo TANG, Baoxian WANG, "Generating Accurate Candidate Windows by Effective Receptive Field" in IEICE TRANSACTIONS on Fundamentals,
vol. E102-A, no. 12, pp. 1925-1927, December 2019, doi: 10.1587/transfun.E102.A.1925.
Abstract: Towards involving the convolutional neural networks into the object detection field, many computer vision tasks have achieved favorable successes. In order to adapt targets with various scales, deep feature pyramid is widely used, since the traditional object detection methods detect different objects in Gaussian image pyramid. However, due to the mismatching between the anchors and the feature distributions of targets, the accurate detection for targets with various scales is still a challenge. Considering the differences between the theoretical receptive field and effective receptive field, we propose a novel anchor generation method, which takes the effective receptive field as the standard. The proposed method is evaluated on the PASCAL VOC dataset and shows the favorable results.
URL: https://global.ieice.org/en_transactions/fundamentals/10.1587/transfun.E102.A.1925/_p
부
@ARTICLE{e102-a_12_1925,
author={Baojun ZHAO, Boya ZHAO, Linbo TANG, Baoxian WANG, },
journal={IEICE TRANSACTIONS on Fundamentals},
title={Generating Accurate Candidate Windows by Effective Receptive Field},
year={2019},
volume={E102-A},
number={12},
pages={1925-1927},
abstract={Towards involving the convolutional neural networks into the object detection field, many computer vision tasks have achieved favorable successes. In order to adapt targets with various scales, deep feature pyramid is widely used, since the traditional object detection methods detect different objects in Gaussian image pyramid. However, due to the mismatching between the anchors and the feature distributions of targets, the accurate detection for targets with various scales is still a challenge. Considering the differences between the theoretical receptive field and effective receptive field, we propose a novel anchor generation method, which takes the effective receptive field as the standard. The proposed method is evaluated on the PASCAL VOC dataset and shows the favorable results.},
keywords={},
doi={10.1587/transfun.E102.A.1925},
ISSN={1745-1337},
month={December},}
부
TY - JOUR
TI - Generating Accurate Candidate Windows by Effective Receptive Field
T2 - IEICE TRANSACTIONS on Fundamentals
SP - 1925
EP - 1927
AU - Baojun ZHAO
AU - Boya ZHAO
AU - Linbo TANG
AU - Baoxian WANG
PY - 2019
DO - 10.1587/transfun.E102.A.1925
JO - IEICE TRANSACTIONS on Fundamentals
SN - 1745-1337
VL - E102-A
IS - 12
JA - IEICE TRANSACTIONS on Fundamentals
Y1 - December 2019
AB - Towards involving the convolutional neural networks into the object detection field, many computer vision tasks have achieved favorable successes. In order to adapt targets with various scales, deep feature pyramid is widely used, since the traditional object detection methods detect different objects in Gaussian image pyramid. However, due to the mismatching between the anchors and the feature distributions of targets, the accurate detection for targets with various scales is still a challenge. Considering the differences between the theoretical receptive field and effective receptive field, we propose a novel anchor generation method, which takes the effective receptive field as the standard. The proposed method is evaluated on the PASCAL VOC dataset and shows the favorable results.
ER -