SSD-EMB: An Improved SSD Using Enhanced Feature Map Block for Object Detection
The development of deep learning has achieved great success in object detection, but small object detection is still a difficult and challenging task in computer vision. To address the problem, we propose an improved single-shot multibox detector (SSD) using enhanced feature map blocks (SSD-EMB). Th...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-04-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/21/8/2842 |
| _version_ | 1797537314194325504 |
|---|---|
| author | Hong-Tae Choi Ho-Jun Lee Hoon Kang Sungwook Yu Ho-Hyun Park |
| author_facet | Hong-Tae Choi Ho-Jun Lee Hoon Kang Sungwook Yu Ho-Hyun Park |
| author_sort | Hong-Tae Choi |
| collection | DOAJ |
| description | The development of deep learning has achieved great success in object detection, but small object detection is still a difficult and challenging task in computer vision. To address the problem, we propose an improved single-shot multibox detector (SSD) using enhanced feature map blocks (SSD-EMB). The enhanced feature map block (EMB) consists of attention stream and feature map concatenation stream. The attention stream allows the proposed model to focus on the object regions rather than background owing to channel averaging and the effectiveness of the normalization. The feature map concatenation stream provides additional semantic information to the model without degrading the detection speed. By combining the output of these two streams, the enhanced feature map, which improves the detection of a small object, is generated. Experimental results show that the proposed model has high accuracy in small object detection. The proposed model not only achieves good detection accuracy, but also has a good detection speed. The SSD-EMB achieved a mean average precision (mAP) of 80.4% on the PASCAL VOC 2007 dataset at 30 frames per second on an RTX 2080Ti graphics processing unit, an mAP of 79.9% on the VOC 2012 dataset, and an mAP of 26.6% on the MS COCO dataset. |
| first_indexed | 2024-03-10T12:13:21Z |
| format | Article |
| id | doaj.art-50417d5e6b2447fd86dc46ca30bc03b0 |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-03-10T12:13:21Z |
| publishDate | 2021-04-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-50417d5e6b2447fd86dc46ca30bc03b02023-11-21T16:02:01ZengMDPI AGSensors1424-82202021-04-01218284210.3390/s21082842SSD-EMB: An Improved SSD Using Enhanced Feature Map Block for Object DetectionHong-Tae Choi0Ho-Jun Lee1Hoon Kang2Sungwook Yu3Ho-Hyun Park4School of Electrical and Electronics Engineering, Chung-Ang University, Seoul 06974, KoreaSchool of Electrical and Electronics Engineering, Chung-Ang University, Seoul 06974, KoreaSchool of Electrical and Electronics Engineering, Chung-Ang University, Seoul 06974, KoreaSchool of Electrical and Electronics Engineering, Chung-Ang University, Seoul 06974, KoreaSchool of Electrical and Electronics Engineering, Chung-Ang University, Seoul 06974, KoreaThe development of deep learning has achieved great success in object detection, but small object detection is still a difficult and challenging task in computer vision. To address the problem, we propose an improved single-shot multibox detector (SSD) using enhanced feature map blocks (SSD-EMB). The enhanced feature map block (EMB) consists of attention stream and feature map concatenation stream. The attention stream allows the proposed model to focus on the object regions rather than background owing to channel averaging and the effectiveness of the normalization. The feature map concatenation stream provides additional semantic information to the model without degrading the detection speed. By combining the output of these two streams, the enhanced feature map, which improves the detection of a small object, is generated. Experimental results show that the proposed model has high accuracy in small object detection. The proposed model not only achieves good detection accuracy, but also has a good detection speed. The SSD-EMB achieved a mean average precision (mAP) of 80.4% on the PASCAL VOC 2007 dataset at 30 frames per second on an RTX 2080Ti graphics processing unit, an mAP of 79.9% on the VOC 2012 dataset, and an mAP of 26.6% on the MS COCO dataset.https://www.mdpi.com/1424-8220/21/8/2842object detectionSSDattention mechanismfeature map concatenation |
| spellingShingle | Hong-Tae Choi Ho-Jun Lee Hoon Kang Sungwook Yu Ho-Hyun Park SSD-EMB: An Improved SSD Using Enhanced Feature Map Block for Object Detection Sensors object detection SSD attention mechanism feature map concatenation |
| title | SSD-EMB: An Improved SSD Using Enhanced Feature Map Block for Object Detection |
| title_full | SSD-EMB: An Improved SSD Using Enhanced Feature Map Block for Object Detection |
| title_fullStr | SSD-EMB: An Improved SSD Using Enhanced Feature Map Block for Object Detection |
| title_full_unstemmed | SSD-EMB: An Improved SSD Using Enhanced Feature Map Block for Object Detection |
| title_short | SSD-EMB: An Improved SSD Using Enhanced Feature Map Block for Object Detection |
| title_sort | ssd emb an improved ssd using enhanced feature map block for object detection |
| topic | object detection SSD attention mechanism feature map concatenation |
| url | https://www.mdpi.com/1424-8220/21/8/2842 |
| work_keys_str_mv | AT hongtaechoi ssdembanimprovedssdusingenhancedfeaturemapblockforobjectdetection AT hojunlee ssdembanimprovedssdusingenhancedfeaturemapblockforobjectdetection AT hoonkang ssdembanimprovedssdusingenhancedfeaturemapblockforobjectdetection AT sungwookyu ssdembanimprovedssdusingenhancedfeaturemapblockforobjectdetection AT hohyunpark ssdembanimprovedssdusingenhancedfeaturemapblockforobjectdetection |