R<sup>2</sup>FA-Det: Delving into High-Quality Rotatable Boxes for Ship Detection in SAR Images
Recently, convolutional neural network (CNN)-based methods have been extensively explored for ship detection in synthetic aperture radar (SAR) images due to their powerful feature representation abilities. However, there are still several obstacles hindering the development. First, ships appear in v...
Main Authors: | , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2020-06-01
|
Series: | Remote Sensing |
Subjects: | |
Online Access: | https://www.mdpi.com/2072-4292/12/12/2031 |
id |
doaj-83d4493e38d7473d8815c8def0f852d4 |
---|---|
record_format |
Article |
spelling |
doaj-83d4493e38d7473d8815c8def0f852d42020-11-25T02:58:50ZengMDPI AGRemote Sensing2072-42922020-06-01122031203110.3390/rs12122031R<sup>2</sup>FA-Det: Delving into High-Quality Rotatable Boxes for Ship Detection in SAR ImagesShiqi Chen0Jun Zhang1Ronghui Zhan2Science and Technology on Automatic Target Recognition Laboratory, National University of Defense Technology, Changsha 410073, ChinaScience and Technology on Automatic Target Recognition Laboratory, National University of Defense Technology, Changsha 410073, ChinaScience and Technology on Automatic Target Recognition Laboratory, National University of Defense Technology, Changsha 410073, ChinaRecently, convolutional neural network (CNN)-based methods have been extensively explored for ship detection in synthetic aperture radar (SAR) images due to their powerful feature representation abilities. However, there are still several obstacles hindering the development. First, ships appear in various scenarios, which makes it difficult to exclude the disruption of the cluttered background. Second, it becomes more complicated to precisely locate the targets with large aspect ratios, arbitrary orientations and dense distributions. Third, the trade-off between accurate localization and improved detection efficiency needs to be considered. To address these issues, this paper presents a rotate refined feature alignment detector (R<inline-formula> <math display="inline"> <semantics> <msup> <mrow></mrow> <mn>2</mn> </msup> </semantics> </math> </inline-formula>FA-Det), which ingeniously balances the quality of bounding box prediction and the high speed of the single-stage framework. Specifically, first, we devise a lightweight non-local attention module and embed it into the stem network. The recalibration of features not only strengthens the object-related features yet adequately suppresses the background interference. In addition, both forms of anchors are integrated into our modified anchor mechanism and thus can enable better representation of densely arranged targets with less computation burden. Furthermore, considering the shortcoming of the feature misalignment existing in the cascaded refinement scheme, a feature-guided alignment module which encodes both the position and shape information of current refined anchors into the feature points is adopted. Extensive experimental validations on two SAR ship datasets are performed and the results demonstrate that our algorithm has higher accuracy with faster speed than some state-of-the-art methods.https://www.mdpi.com/2072-4292/12/12/2031attention modulecascade architecturefeature guided alignment modulemodified anchor mechanismsingle-stage detectorsynthetic aperture radar (SAR) ship detection |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Shiqi Chen Jun Zhang Ronghui Zhan |
spellingShingle |
Shiqi Chen Jun Zhang Ronghui Zhan R<sup>2</sup>FA-Det: Delving into High-Quality Rotatable Boxes for Ship Detection in SAR Images Remote Sensing attention module cascade architecture feature guided alignment module modified anchor mechanism single-stage detector synthetic aperture radar (SAR) ship detection |
author_facet |
Shiqi Chen Jun Zhang Ronghui Zhan |
author_sort |
Shiqi Chen |
title |
R<sup>2</sup>FA-Det: Delving into High-Quality Rotatable Boxes for Ship Detection in SAR Images |
title_short |
R<sup>2</sup>FA-Det: Delving into High-Quality Rotatable Boxes for Ship Detection in SAR Images |
title_full |
R<sup>2</sup>FA-Det: Delving into High-Quality Rotatable Boxes for Ship Detection in SAR Images |
title_fullStr |
R<sup>2</sup>FA-Det: Delving into High-Quality Rotatable Boxes for Ship Detection in SAR Images |
title_full_unstemmed |
R<sup>2</sup>FA-Det: Delving into High-Quality Rotatable Boxes for Ship Detection in SAR Images |
title_sort |
r<sup>2</sup>fa-det: delving into high-quality rotatable boxes for ship detection in sar images |
publisher |
MDPI AG |
series |
Remote Sensing |
issn |
2072-4292 |
publishDate |
2020-06-01 |
description |
Recently, convolutional neural network (CNN)-based methods have been extensively explored for ship detection in synthetic aperture radar (SAR) images due to their powerful feature representation abilities. However, there are still several obstacles hindering the development. First, ships appear in various scenarios, which makes it difficult to exclude the disruption of the cluttered background. Second, it becomes more complicated to precisely locate the targets with large aspect ratios, arbitrary orientations and dense distributions. Third, the trade-off between accurate localization and improved detection efficiency needs to be considered. To address these issues, this paper presents a rotate refined feature alignment detector (R<inline-formula> <math display="inline"> <semantics> <msup> <mrow></mrow> <mn>2</mn> </msup> </semantics> </math> </inline-formula>FA-Det), which ingeniously balances the quality of bounding box prediction and the high speed of the single-stage framework. Specifically, first, we devise a lightweight non-local attention module and embed it into the stem network. The recalibration of features not only strengthens the object-related features yet adequately suppresses the background interference. In addition, both forms of anchors are integrated into our modified anchor mechanism and thus can enable better representation of densely arranged targets with less computation burden. Furthermore, considering the shortcoming of the feature misalignment existing in the cascaded refinement scheme, a feature-guided alignment module which encodes both the position and shape information of current refined anchors into the feature points is adopted. Extensive experimental validations on two SAR ship datasets are performed and the results demonstrate that our algorithm has higher accuracy with faster speed than some state-of-the-art methods. |
topic |
attention module cascade architecture feature guided alignment module modified anchor mechanism single-stage detector synthetic aperture radar (SAR) ship detection |
url |
https://www.mdpi.com/2072-4292/12/12/2031 |
work_keys_str_mv |
AT shiqichen rsup2supfadetdelvingintohighqualityrotatableboxesforshipdetectioninsarimages AT junzhang rsup2supfadetdelvingintohighqualityrotatableboxesforshipdetectioninsarimages AT ronghuizhan rsup2supfadetdelvingintohighqualityrotatableboxesforshipdetectioninsarimages |
_version_ |
1724704875153981440 |