Ghostformer: A GhostNet-Based Two-Stage Transformer for Small Object Detection
In this paper, we propose a novel two-stage transformer with GhostNet, which improves the performance of the small object detection task. Specifically, based on the original Deformable Transformers for End-to-End Object Detection (deformable DETR), we chose GhostNet as the backbone to extract featur...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-09-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/22/18/6939 |
| _version_ | 1827656870446235648 |
|---|---|
| author | Sijia Li Furkat Sultonov Jamshid Tursunboev Jun-Hyun Park Sangseok Yun Jae-Mo Kang |
| author_facet | Sijia Li Furkat Sultonov Jamshid Tursunboev Jun-Hyun Park Sangseok Yun Jae-Mo Kang |
| author_sort | Sijia Li |
| collection | DOAJ |
| description | In this paper, we propose a novel two-stage transformer with GhostNet, which improves the performance of the small object detection task. Specifically, based on the original Deformable Transformers for End-to-End Object Detection (deformable DETR), we chose GhostNet as the backbone to extract features, since it is better suited for an efficient feature extraction. Furthermore, at the target detection stage, we selected the 300 best bounding box results as <i>regional proposals,</i> which were subsequently set as primary object queries of the decoder layer. Finally, in the decoder layer, we optimized and modified the queries to increase the target accuracy. In order to validate the performance of the proposed model, we adopted a widely used COCO 2017 dataset. Extensive experiments demonstrated that the proposed scheme yielded a higher average precision (AP) score in detecting small objects than the existing deformable DETR model. |
| first_indexed | 2024-03-09T22:34:24Z |
| format | Article |
| id | doaj.art-fdabcbbf5d09427fb372899d023eaec6 |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-03-09T22:34:24Z |
| publishDate | 2022-09-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-fdabcbbf5d09427fb372899d023eaec62023-11-23T18:51:49ZengMDPI AGSensors1424-82202022-09-012218693910.3390/s22186939Ghostformer: A GhostNet-Based Two-Stage Transformer for Small Object DetectionSijia Li0Furkat Sultonov1Jamshid Tursunboev2Jun-Hyun Park3Sangseok Yun4Jae-Mo Kang5Department of Artificial Intelligence, Kyungpook National University, Daegu 41566, KoreaDepartment of Artificial Intelligence, Kyungpook National University, Daegu 41566, KoreaDepartment of Artificial Intelligence, Kyungpook National University, Daegu 41566, KoreaDepartment of Artificial Intelligence, Kyungpook National University, Daegu 41566, KoreaDepartment of Information and Communications Engineering, Pukyong National University, Busan 48513, KoreaDepartment of Artificial Intelligence, Kyungpook National University, Daegu 41566, KoreaIn this paper, we propose a novel two-stage transformer with GhostNet, which improves the performance of the small object detection task. Specifically, based on the original Deformable Transformers for End-to-End Object Detection (deformable DETR), we chose GhostNet as the backbone to extract features, since it is better suited for an efficient feature extraction. Furthermore, at the target detection stage, we selected the 300 best bounding box results as <i>regional proposals,</i> which were subsequently set as primary object queries of the decoder layer. Finally, in the decoder layer, we optimized and modified the queries to increase the target accuracy. In order to validate the performance of the proposed model, we adopted a widely used COCO 2017 dataset. Extensive experiments demonstrated that the proposed scheme yielded a higher average precision (AP) score in detecting small objects than the existing deformable DETR model.https://www.mdpi.com/1424-8220/22/18/6939small object detectionGhostNet<i>regional proposals</i>two-stage transformer |
| spellingShingle | Sijia Li Furkat Sultonov Jamshid Tursunboev Jun-Hyun Park Sangseok Yun Jae-Mo Kang Ghostformer: A GhostNet-Based Two-Stage Transformer for Small Object Detection Sensors small object detection GhostNet <i>regional proposals</i> two-stage transformer |
| title | Ghostformer: A GhostNet-Based Two-Stage Transformer for Small Object Detection |
| title_full | Ghostformer: A GhostNet-Based Two-Stage Transformer for Small Object Detection |
| title_fullStr | Ghostformer: A GhostNet-Based Two-Stage Transformer for Small Object Detection |
| title_full_unstemmed | Ghostformer: A GhostNet-Based Two-Stage Transformer for Small Object Detection |
| title_short | Ghostformer: A GhostNet-Based Two-Stage Transformer for Small Object Detection |
| title_sort | ghostformer a ghostnet based two stage transformer for small object detection |
| topic | small object detection GhostNet <i>regional proposals</i> two-stage transformer |
| url | https://www.mdpi.com/1424-8220/22/18/6939 |
| work_keys_str_mv | AT sijiali ghostformeraghostnetbasedtwostagetransformerforsmallobjectdetection AT furkatsultonov ghostformeraghostnetbasedtwostagetransformerforsmallobjectdetection AT jamshidtursunboev ghostformeraghostnetbasedtwostagetransformerforsmallobjectdetection AT junhyunpark ghostformeraghostnetbasedtwostagetransformerforsmallobjectdetection AT sangseokyun ghostformeraghostnetbasedtwostagetransformerforsmallobjectdetection AT jaemokang ghostformeraghostnetbasedtwostagetransformerforsmallobjectdetection |