Dynamic Anchor: A Feature-Guided Anchor Strategy for Object Detection
The majority of modern object detectors rely on a set of pre-defined anchor boxes, which enhances detection performance dramatically. Nevertheless, the pre-defined anchor strategy suffers some drawbacks, especially the complex hyper-parameters of anchors, seriously affecting detection performance. I...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-05-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/12/10/4897 |
| _version_ | 1797501827266117632 |
|---|---|
| author | Xing Liu Huai-Xin Chen Bi-Yuan Liu |
| author_facet | Xing Liu Huai-Xin Chen Bi-Yuan Liu |
| author_sort | Xing Liu |
| collection | DOAJ |
| description | The majority of modern object detectors rely on a set of pre-defined anchor boxes, which enhances detection performance dramatically. Nevertheless, the pre-defined anchor strategy suffers some drawbacks, especially the complex hyper-parameters of anchors, seriously affecting detection performance. In this paper, we propose a feature-guided anchor generation method named dynamic anchor. Dynamic anchor mainly includes two structures: the anchor generator and the feature enhancement module. The anchor generator leverages semantic features to predict optimized anchor shapes at the locations where the objects are likely to exist in the feature maps; by converting the predicted shape maps into location offsets, the feature enhancement module uses the high-quality anchors to improve detection performance. Compared with the hand-designed anchor scheme, dynamic anchor discards all pre-defined boxes and avoids complex hyper-parameters. In addition, only one anchor box is predicted for each location, which dramatically reduces calculation. With ResNet-50 and ResNet-101 as the backbone of the one-stage detector RetinaNet, dynamic anchor achieved 2.1 AP and 1.0 AP gains, respectively. The proposed dynamic anchor strategy can be easily integrated into the anchor-based detectors to replace the traditional pre-defined anchor scheme. |
| first_indexed | 2024-03-10T03:24:12Z |
| format | Article |
| id | doaj.art-2e48000b06dd423bbca3abd8030ecea4 |
| institution | Directory Open Access Journal |
| issn | 2076-3417 |
| language | English |
| last_indexed | 2024-03-10T03:24:12Z |
| publishDate | 2022-05-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj.art-2e48000b06dd423bbca3abd8030ecea42023-11-23T09:54:47ZengMDPI AGApplied Sciences2076-34172022-05-011210489710.3390/app12104897Dynamic Anchor: A Feature-Guided Anchor Strategy for Object DetectionXing Liu0Huai-Xin Chen1Bi-Yuan Liu2School of Resources and Environment, University of Electronic Science and Technology of China, Chengdu 611731, ChinaSchool of Resources and Environment, University of Electronic Science and Technology of China, Chengdu 611731, ChinaSchool of Resources and Environment, University of Electronic Science and Technology of China, Chengdu 611731, ChinaThe majority of modern object detectors rely on a set of pre-defined anchor boxes, which enhances detection performance dramatically. Nevertheless, the pre-defined anchor strategy suffers some drawbacks, especially the complex hyper-parameters of anchors, seriously affecting detection performance. In this paper, we propose a feature-guided anchor generation method named dynamic anchor. Dynamic anchor mainly includes two structures: the anchor generator and the feature enhancement module. The anchor generator leverages semantic features to predict optimized anchor shapes at the locations where the objects are likely to exist in the feature maps; by converting the predicted shape maps into location offsets, the feature enhancement module uses the high-quality anchors to improve detection performance. Compared with the hand-designed anchor scheme, dynamic anchor discards all pre-defined boxes and avoids complex hyper-parameters. In addition, only one anchor box is predicted for each location, which dramatically reduces calculation. With ResNet-50 and ResNet-101 as the backbone of the one-stage detector RetinaNet, dynamic anchor achieved 2.1 AP and 1.0 AP gains, respectively. The proposed dynamic anchor strategy can be easily integrated into the anchor-based detectors to replace the traditional pre-defined anchor scheme.https://www.mdpi.com/2076-3417/12/10/4897object detectoranchor generation strategyfeature enhancementRetinaNet |
| spellingShingle | Xing Liu Huai-Xin Chen Bi-Yuan Liu Dynamic Anchor: A Feature-Guided Anchor Strategy for Object Detection Applied Sciences object detector anchor generation strategy feature enhancement RetinaNet |
| title | Dynamic Anchor: A Feature-Guided Anchor Strategy for Object Detection |
| title_full | Dynamic Anchor: A Feature-Guided Anchor Strategy for Object Detection |
| title_fullStr | Dynamic Anchor: A Feature-Guided Anchor Strategy for Object Detection |
| title_full_unstemmed | Dynamic Anchor: A Feature-Guided Anchor Strategy for Object Detection |
| title_short | Dynamic Anchor: A Feature-Guided Anchor Strategy for Object Detection |
| title_sort | dynamic anchor a feature guided anchor strategy for object detection |
| topic | object detector anchor generation strategy feature enhancement RetinaNet |
| url | https://www.mdpi.com/2076-3417/12/10/4897 |
| work_keys_str_mv | AT xingliu dynamicanchorafeatureguidedanchorstrategyforobjectdetection AT huaixinchen dynamicanchorafeatureguidedanchorstrategyforobjectdetection AT biyuanliu dynamicanchorafeatureguidedanchorstrategyforobjectdetection |