DDL R-CNN: Dynamic Direction Learning R-CNN for Rotated Object Detection
Current remote sensing (RS) detectors often rely on predefined anchor boxes with fixed angles to handle the multi-directional variations of targets. This approach makes it challenging to accurately select regions of interest and extract features that align with the direction of the targets. Most exi...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-01-01
|
| Series: | Algorithms |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1999-4893/18/1/21 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832589395069239296 |
|---|---|
| author | Weixian Su Donglin Jing |
| author_facet | Weixian Su Donglin Jing |
| author_sort | Weixian Su |
| collection | DOAJ |
| description | Current remote sensing (RS) detectors often rely on predefined anchor boxes with fixed angles to handle the multi-directional variations of targets. This approach makes it challenging to accurately select regions of interest and extract features that align with the direction of the targets. Most existing regression methods also adopt angle regression to match the attributes of remote sensing detectors. Due to the inconsistent regression direction and massive anchor boxes with a high aspect ratio, the extracted target features change greatly, the loss function changes drastically, and the training is unstable. However, existing RS detectors and regression techniques have not been able to effectively balance the precision of directional feature extraction with the complexity of the models. To address these challenges, this paper introduces a novel approach known as Dynamic Direction Learning R-CNN (DDL R-CNN), which comprises a dynamic direction learning (DDL) module and a boundary center region offset generation network (BC-ROPN). The DDL module pre-extracts the directional features of targets to provide a coarse estimation of their angles and the corresponding weights. This information is used to generate rotationally aligned anchor boxes that better model the directional features of the targets. BC-ROPN represents an innovative method for anchor box regression. It utilizes the central features of the maximum bounding rectangle’s width and height, along with the coarse angle estimation and weights derived from DDL module, to refine the orientation of the anchor box. Our method has been proven to surpass existing rotating detection networks in extensive testing across two widely used remote sensing detection datasets, namely UCAS-AOD and HRSC2016. |
| format | Article |
| id | doaj-art-440432c79ed943e4809d25e1f054a021 |
| institution | Kabale University |
| issn | 1999-4893 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Algorithms |
| spelling | doaj-art-440432c79ed943e4809d25e1f054a0212025-01-24T13:17:30ZengMDPI AGAlgorithms1999-48932025-01-011812110.3390/a18010021DDL R-CNN: Dynamic Direction Learning R-CNN for Rotated Object DetectionWeixian Su0Donglin Jing1Faculty of Engineering, University of Hong Kong, Hong Kong 999077, ChinaThe School of Information and Electronics, Beijing Institute of Technology, Beijing 100081, ChinaCurrent remote sensing (RS) detectors often rely on predefined anchor boxes with fixed angles to handle the multi-directional variations of targets. This approach makes it challenging to accurately select regions of interest and extract features that align with the direction of the targets. Most existing regression methods also adopt angle regression to match the attributes of remote sensing detectors. Due to the inconsistent regression direction and massive anchor boxes with a high aspect ratio, the extracted target features change greatly, the loss function changes drastically, and the training is unstable. However, existing RS detectors and regression techniques have not been able to effectively balance the precision of directional feature extraction with the complexity of the models. To address these challenges, this paper introduces a novel approach known as Dynamic Direction Learning R-CNN (DDL R-CNN), which comprises a dynamic direction learning (DDL) module and a boundary center region offset generation network (BC-ROPN). The DDL module pre-extracts the directional features of targets to provide a coarse estimation of their angles and the corresponding weights. This information is used to generate rotationally aligned anchor boxes that better model the directional features of the targets. BC-ROPN represents an innovative method for anchor box regression. It utilizes the central features of the maximum bounding rectangle’s width and height, along with the coarse angle estimation and weights derived from DDL module, to refine the orientation of the anchor box. Our method has been proven to surpass existing rotating detection networks in extensive testing across two widely used remote sensing detection datasets, namely UCAS-AOD and HRSC2016.https://www.mdpi.com/1999-4893/18/1/21remote sensing detectionanchor box regressiondirectional feature extractionhigh-aspect-ratio detection |
| spellingShingle | Weixian Su Donglin Jing DDL R-CNN: Dynamic Direction Learning R-CNN for Rotated Object Detection Algorithms remote sensing detection anchor box regression directional feature extraction high-aspect-ratio detection |
| title | DDL R-CNN: Dynamic Direction Learning R-CNN for Rotated Object Detection |
| title_full | DDL R-CNN: Dynamic Direction Learning R-CNN for Rotated Object Detection |
| title_fullStr | DDL R-CNN: Dynamic Direction Learning R-CNN for Rotated Object Detection |
| title_full_unstemmed | DDL R-CNN: Dynamic Direction Learning R-CNN for Rotated Object Detection |
| title_short | DDL R-CNN: Dynamic Direction Learning R-CNN for Rotated Object Detection |
| title_sort | ddl r cnn dynamic direction learning r cnn for rotated object detection |
| topic | remote sensing detection anchor box regression directional feature extraction high-aspect-ratio detection |
| url | https://www.mdpi.com/1999-4893/18/1/21 |
| work_keys_str_mv | AT weixiansu ddlrcnndynamicdirectionlearningrcnnforrotatedobjectdetection AT donglinjing ddlrcnndynamicdirectionlearningrcnnforrotatedobjectdetection |