Learning Deep Relationship for Object Detection

Nuo XU; Chunlei HUO

doi:10.1587/transinf.2017EDL8131

Learning Deep Relationship for Object Detection

Nuo XU, Chunlei HUO

Full Text Views

0

Share
Cite this

Summary :

Object detection has been a hot topic of image processing, computer vision and pattern recognition. In recent years, training a model from labeled images using machine learning technique becomes popular. However, the relationship between training samples is usually ignored by existing approaches. To address this problem, a novel approach is proposed, which trains Siamese convolutional neural network on feature pairs and finely tunes the network driven by a small amount of training samples. Since the proposed method considers not only the discriminative information between objects and background, but also the relationship between intraclass features, it outperforms the state-of-arts on real images.

Publication: IEICE TRANSACTIONS on Information Vol.E101-D No.1 pp.273-276

Publication Date: 2018/01/01

Publicized: 2017/09/28

Online ISSN: 1745-1361

DOI: 10.1587/transinf.2017EDL8131

Type of Manuscript: LETTER

Category: Image Recognition, Computer Vision

Authors

Nuo XU
Chinese Academy of Sciences
Chunlei HUO
Chinese Academy of Sciences

Keyword

object detection, Siamese convolutional neural network, remote sensing images, relationship learning

Cite this

Copy

Nuo XU, Chunlei HUO, "Learning Deep Relationship for Object Detection" in IEICE TRANSACTIONS on Information, vol. E101-D, no. 1, pp. 273-276, January 2018, doi: 10.1587/transinf.2017EDL8131.
Abstract: Object detection has been a hot topic of image processing, computer vision and pattern recognition. In recent years, training a model from labeled images using machine learning technique becomes popular. However, the relationship between training samples is usually ignored by existing approaches. To address this problem, a novel approach is proposed, which trains Siamese convolutional neural network on feature pairs and finely tunes the network driven by a small amount of training samples. Since the proposed method considers not only the discriminative information between objects and background, but also the relationship between intraclass features, it outperforms the state-of-arts on real images.
URL: https://globals.ieice.org/en_transactions/information/10.1587/transinf.2017EDL8131/_p

Copy

@ARTICLE{e101-d_1_273,
author={Nuo XU, Chunlei HUO, },
journal={IEICE TRANSACTIONS on Information},
title={Learning Deep Relationship for Object Detection},
year={2018},
volume={E101-D},
number={1},
pages={273-276},
abstract={Object detection has been a hot topic of image processing, computer vision and pattern recognition. In recent years, training a model from labeled images using machine learning technique becomes popular. However, the relationship between training samples is usually ignored by existing approaches. To address this problem, a novel approach is proposed, which trains Siamese convolutional neural network on feature pairs and finely tunes the network driven by a small amount of training samples. Since the proposed method considers not only the discriminative information between objects and background, but also the relationship between intraclass features, it outperforms the state-of-arts on real images.},
keywords={},
doi={10.1587/transinf.2017EDL8131},
ISSN={1745-1361},
month={January},}

Copy

TY - JOUR
TI - Learning Deep Relationship for Object Detection
T2 - IEICE TRANSACTIONS on Information
SP - 273
EP - 276
AU - Nuo XU
AU - Chunlei HUO
PY - 2018
DO - 10.1587/transinf.2017EDL8131
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E101-D
IS - 1
JA - IEICE TRANSACTIONS on Information
Y1 - January 2018
AB - Object detection has been a hot topic of image processing, computer vision and pattern recognition. In recent years, training a model from labeled images using machine learning technique becomes popular. However, the relationship between training samples is usually ignored by existing approaches. To address this problem, a novel approach is proposed, which trains Siamese convolutional neural network on feature pairs and finely tunes the network driven by a small amount of training samples. Since the proposed method considers not only the discriminative information between objects and background, but also the relationship between intraclass features, it outperforms the state-of-arts on real images.
ER -