The Object detection by the combination of generic roi extractor and dynamic R-CNN with side-aware boundary localization in aerial images
Unmanned Aerial Vehicles (UAVs) have recently gained popularity due to their simplicity and effectiveness in traffic monitoring and potential for rapid delivery, and rescue support. Moreover, UAVs have been employed as a supporting machine in data collection for object detection tasks, in particula...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
CAN THO UNIVERSITY PUBLISHING HOUSE
2023-03-01
|
| Series: | Can Tho University Journal of Science |
| Subjects: | |
| Online Access: | https://ctujs.ctu.edu.vn/index.php/ctujs/article/view/477 |
| _version_ | 1797847657726607360 |
|---|---|
| author | Nguyen Bao Tran Tan Tai Pham Cao Doanh Bui Nguyen D. Vo Khang Nguyen |
| author_facet | Nguyen Bao Tran Tan Tai Pham Cao Doanh Bui Nguyen D. Vo Khang Nguyen |
| author_sort | Nguyen Bao Tran |
| collection | DOAJ |
| description |
Unmanned Aerial Vehicles (UAVs) have recently gained popularity due to their simplicity and effectiveness in traffic monitoring and potential for rapid delivery, and rescue support. Moreover, UAVs have been employed as a supporting machine in data collection for object detection tasks, in particular vehicle detection tasks in object recognition. Although vehicle identification is a tough problem, many of its challenges have recently been overcome by two-stage approaches such as Faster R-CNN, one of the most successful vehicle detectors. However, many critical problems still remain, such as partial occlusion, object truncation, object multi-angle rotation, etc. In this paper, we combine the Generic RoI Extractor (GroIE) method with Dynamic R-CNN and Side-aware Boundary Localization (SABL) for both testing and evaluation on a challenging dataset XDUAV. Overall, 4344 images in the XDUAV dataset, divided into 3 subsets: 3485 training images, 869 testing images and 869 validating images were used. These consisted of six object classes: 33841 “car”; 2690 “bus”; 2848 “truck”; 173 “tanker”; 6656 “motor” and 2024 “bicycle”. With the ResNet-101 backbone, our approach showed competitive results compared with the original GRoIE method, surpassed by 1.2% on mAP score and by about 2% on most classes AP scores, except for the class 'tanker'.
|
| first_indexed | 2024-04-09T18:16:06Z |
| format | Article |
| id | doaj.art-d4b474d60bbc4c7c906d8f1f4d6f4c43 |
| institution | Directory Open Access Journal |
| issn | 2615-9422 2815-5602 |
| language | English |
| last_indexed | 2024-04-09T18:16:06Z |
| publishDate | 2023-03-01 |
| publisher | CAN THO UNIVERSITY PUBLISHING HOUSE |
| record_format | Article |
| series | Can Tho University Journal of Science |
| spelling | doaj.art-d4b474d60bbc4c7c906d8f1f4d6f4c432023-04-13T04:46:23ZengCAN THO UNIVERSITY PUBLISHING HOUSECan Tho University Journal of Science2615-94222815-56022023-03-0115110.22144/ctu.jen.2023.006The Object detection by the combination of generic roi extractor and dynamic R-CNN with side-aware boundary localization in aerial imagesNguyen Bao Tran0Tan Tai Pham1Cao Doanh Bui2Nguyen D. VoKhang NguyenUniversity of Information Technology - VNUHCMUniversity of Information Technology - VNUHCMUniversity of Information Technology - VNUHCM Unmanned Aerial Vehicles (UAVs) have recently gained popularity due to their simplicity and effectiveness in traffic monitoring and potential for rapid delivery, and rescue support. Moreover, UAVs have been employed as a supporting machine in data collection for object detection tasks, in particular vehicle detection tasks in object recognition. Although vehicle identification is a tough problem, many of its challenges have recently been overcome by two-stage approaches such as Faster R-CNN, one of the most successful vehicle detectors. However, many critical problems still remain, such as partial occlusion, object truncation, object multi-angle rotation, etc. In this paper, we combine the Generic RoI Extractor (GroIE) method with Dynamic R-CNN and Side-aware Boundary Localization (SABL) for both testing and evaluation on a challenging dataset XDUAV. Overall, 4344 images in the XDUAV dataset, divided into 3 subsets: 3485 training images, 869 testing images and 869 validating images were used. These consisted of six object classes: 33841 “car”; 2690 “bus”; 2848 “truck”; 173 “tanker”; 6656 “motor” and 2024 “bicycle”. With the ResNet-101 backbone, our approach showed competitive results compared with the original GRoIE method, surpassed by 1.2% on mAP score and by about 2% on most classes AP scores, except for the class 'tanker'. https://ctujs.ctu.edu.vn/index.php/ctujs/article/view/477Vehicle detectionobject detectionUAV datasetsXDUAV dataset |
| spellingShingle | Nguyen Bao Tran Tan Tai Pham Cao Doanh Bui Nguyen D. Vo Khang Nguyen The Object detection by the combination of generic roi extractor and dynamic R-CNN with side-aware boundary localization in aerial images Can Tho University Journal of Science Vehicle detection object detection UAV datasets XDUAV dataset |
| title | The Object detection by the combination of generic roi extractor and dynamic R-CNN with side-aware boundary localization in aerial images |
| title_full | The Object detection by the combination of generic roi extractor and dynamic R-CNN with side-aware boundary localization in aerial images |
| title_fullStr | The Object detection by the combination of generic roi extractor and dynamic R-CNN with side-aware boundary localization in aerial images |
| title_full_unstemmed | The Object detection by the combination of generic roi extractor and dynamic R-CNN with side-aware boundary localization in aerial images |
| title_short | The Object detection by the combination of generic roi extractor and dynamic R-CNN with side-aware boundary localization in aerial images |
| title_sort | object detection by the combination of generic roi extractor and dynamic r cnn with side aware boundary localization in aerial images |
| topic | Vehicle detection object detection UAV datasets XDUAV dataset |
| url | https://ctujs.ctu.edu.vn/index.php/ctujs/article/view/477 |
| work_keys_str_mv | AT nguyenbaotran theobjectdetectionbythecombinationofgenericroiextractoranddynamicrcnnwithsideawareboundarylocalizationinaerialimages AT tantaipham theobjectdetectionbythecombinationofgenericroiextractoranddynamicrcnnwithsideawareboundarylocalizationinaerialimages AT caodoanhbui theobjectdetectionbythecombinationofgenericroiextractoranddynamicrcnnwithsideawareboundarylocalizationinaerialimages AT nguyendvo theobjectdetectionbythecombinationofgenericroiextractoranddynamicrcnnwithsideawareboundarylocalizationinaerialimages AT khangnguyen theobjectdetectionbythecombinationofgenericroiextractoranddynamicrcnnwithsideawareboundarylocalizationinaerialimages AT nguyenbaotran objectdetectionbythecombinationofgenericroiextractoranddynamicrcnnwithsideawareboundarylocalizationinaerialimages AT tantaipham objectdetectionbythecombinationofgenericroiextractoranddynamicrcnnwithsideawareboundarylocalizationinaerialimages AT caodoanhbui objectdetectionbythecombinationofgenericroiextractoranddynamicrcnnwithsideawareboundarylocalizationinaerialimages AT nguyendvo objectdetectionbythecombinationofgenericroiextractoranddynamicrcnnwithsideawareboundarylocalizationinaerialimages AT khangnguyen objectdetectionbythecombinationofgenericroiextractoranddynamicrcnnwithsideawareboundarylocalizationinaerialimages |