A Recursive Prediction-Based Feature Enhancement for Small Object Detection
<p>Diagram of RSP−DETR for small object detection. The backbone feature layers adopt SAC; RSP applies two−stage filtering to the decoder’s output to compensate for the lack of information on small targets. Furthermore, the loss function has been improved to enhance the performance of DINO in detecting small objects.</p> "> Figure 2
<p>Switchable Atrous Convolution (SAC).</p> "> Figure 3
<p>Categorical and areal information is utilized for differentiation and mapping onto the backbone.</p> "> Figure 4
<p>Variations in positional offset across different scales of <span class="html-italic">IOU</span>.</p> "> Figure 5
<p>Predictive outcomes of baseline model and RSP−DETR across various scenarios: (<b>a</b>) Ground truth; (<b>b</b>) DINO; (<b>c</b>) RSP−DETR. Red circles indicate missed targets and orange circles represent erroneously detected ones.</p> "> Figure 6
<p>Prediction results from the baseline model and RSP−DETR on the HRSID dataset: (<b>a</b>) Ground truth; (<b>b</b>) DINO; (<b>c</b>) RSP−DETR; red circles denote missed targets, and orange circles indicate erroneously detected targets.</p> ">
Abstract
:1. Introduction
- In our study, we scrutinize DETR’s backbone network aimed at feature extraction, opting to enhance it with a technique called Switchable Atrous Convolution, supplanting the standard 3 × 3 convolutions. This versatile method dynamically adjusts the receptive field and assimilates global context, significantly boosting the ability to detect small objects.
- To mitigate the deficiencies of DETR’s decoder in handling small object detection, we introduce a novel Recursive Small Object Prediction (RSP) module. For implementation, the RSP module initially filters decoder predictions based on classification scores and object area, mapping the sifted tokens back onto the backbone’s low−dimensional feature layers. It then refines these initial results to enable more exhaustive small object feature extraction, resulting in an augmented tally of small object predictions.
- Due to DETR’s lack of consideration for small objects in its loss function, we integrate the Normalized Wasserstein Distance (NWD) metric to refine the loss calculation, and through experimentation, we confirm its effectiveness. NWD loss presents a refined metric for gauging similarity, particularly amongst small objects, enhancing the precision of detection.
2. Related Work
2.1. CNN−Based Object Detection
2.2. Small Object Detection
3. Methods
3.1. Switchable Atrous Convolution
3.2. Recursive Small Object Prediction
3.3. Normalized Wasserstein Distance
4. Experiments
4.1. Datasets
4.2. Implementation Details
4.3. Module Effectiveness Analysis
4.4. Ablation Study
Comparison of Inference Accuracy
4.5. Visualization
5. Conclusions
Author Contributions
Funding
Institutional Review Board Statement
Informed Consent Statement
Data Availability Statement
Conflicts of Interest
References
- Girshick, R. Fast r-cnn. In Proceedings of the IEEE International Conference on Computer Vision, Santiago, Chile, 7–13 December 2015; IEEE: Piscataway, NJ, USA, 2015; pp. 1440–1448. [Google Scholar]
- Liu, W.; Anguelov, D.; Erhan, D.; Szegedy, C.; Reed, S.; Fu, C.-Y.; Berg, A.C. Ssd: Single shot multibox detector. In Proceedings of the Computer Vision–ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, 11–14 October 2016; Proceedings, Part I 14. pp. 21–37. [Google Scholar]
- Redmon, J.; Farhadi, A. YOLO9000: Better, faster, stronger. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, HI, USA, 21–26 June 2017; IEEE: Piscataway, NJ, USA, 2017; pp. 7263–7271. [Google Scholar]
- Ren, S.; He, K.; Girshick, R.; Sun, J. Faster r-cnn: Towards Real-Time Object Detection with Region Proposal Networks. IEEE Trans. Pattern Anal. Mach. Intell. 2017, 39, 1137–1149. [Google Scholar] [CrossRef] [PubMed]
- Carion, N.; Massa, F.; Synnaeve, G.; Usunier, N.; Kirillov, A.; Zagoruyko, S. End-to-end object detection with transformers. In Proceedings of the European Conference on Computer Vision, Glasgow, UK, 23–28 August 2020; Springer International Publishing: Cham, Switzerland, 2020; pp. 213–229. [Google Scholar]
- He, K.; Zhang, X.; Ren, S.; Sun, J. Deep residual learning for image recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, NV, USA, 27–30 June 2016; IEEE: Piscataway, NJ, USA, 2016; pp. 770–778. [Google Scholar]
- Simonyan, K.; Zisserman, A. Very deep convolutional networks for large-scale image recognition. arXiv 2014, arXiv:1409.1556. [Google Scholar]
- Wang, C.-Y.; Liao, H.-Y.M.; Wu, Y.-H.; Chen, P.-Y.; Hsieh, J.-W.; Yeh, I.-H. CSPNet: A new backbone that can enhance learning capability of CNN. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, Seattle, WA, USA, 13–19 June 2020; IEEE: Piscataway, NJ, USA, 2020; pp. 390–391. [Google Scholar]
- Vaswani, A.; Shazeer, N.; Parmar, N.; Uszkoreit, J.; Jones, L.; Gomez, A.N.; Kaiser, Ł.; Polosukhin, I. Attention is all you need. Adv. Neural Inf. Process. Syst. 2017, 30. [Google Scholar]
- Zhu, X.; Su, W.; Lu, L.; Li, B.; Wang, X.; Dai, J. Deformable detr: Deformable transformers for end-to-end object detection. arXiv 2020, arXiv:2010.04159. [Google Scholar]
- Roh, B.; Shin, J.; Shin, W.; Kim, S. Sparse detr: Efficient end-to-end object detection with learnable sparsity. arXiv 2021, arXiv:2111.14330. [Google Scholar]
- Yao, Z.; Ai, J.; Li, B.; Zhang, C. Efficient detr: Improving end-to-end object detector with dense prior. arXiv 2021, arXiv:2104.01318. [Google Scholar]
- Li, F.; Zhang, H.; Liu, S.; Guo, J.; Ni, L.M.; Zhang, L. Dn-detr: Accelerate detr training by introducing query denoising. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, New Orleans, LA, USA, 18–24 June 2022; IEEE: Piscataway, NJ, USA, 2022; pp. 13619–13627. [Google Scholar]
- Chen, G.; Wang, H.; Chen, K.; Li, Z.; Song, Z.; Liu, Y.; Chen, W.; Knoll, A. A survey of the four pillars for small object detection: Multiscale representation, contextual information, super-resolution, and region proposal. IEEE Trans. Syst. Man Cybern. Syst. 2020, 52, 936–953. [Google Scholar] [CrossRef]
- Tian, Z.; Chu, X.; Wang, X.; Wei, X.; Shen, C. Fully convolutional one-stage 3d object detection on lidar range images. Adv. Neural Inf. Process. Syst. 2022, 35, 34899–34911. [Google Scholar]
- Chen, F.; Zhang, H.; Hu, K.; Huang, Y.-k.; Zhu, C.; Savvides, M. Enhanced training of query-based object detection via selective query recollection. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Vancouver, BC, Canada, 17–24 June 2023; IEEE: Piscataway, NJ, USA, 2023; pp. 23756–23765. [Google Scholar]
- Zhang, H.; Li, F.; Liu, S.; Zhang, L.; Su, H.; Zhu, J.; Ni, L.M.; Shum, H.-Y. Dino: Detr with improved denoising anchor boxes for end-to-end object detection. arXiv 2022, arXiv:2203.03605. [Google Scholar]
- Qiao, S.; Chen, L.-C.; Yuille, A. Detectors: Detecting objects with recursive feature pyramid and switchable atrous convolution. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Nashville, TN, USA, 20–25 June 2021; IEEE: Piscataway, NJ, USA, 2021; pp. 10213–10224. [Google Scholar]
- Yu, F.; Koltun, V. Multi-scale context aggregation by dilated convolutions. arXiv 2015, arXiv:1511.07122. [Google Scholar]
- Jang, D.-H.; Chu, S.; Kim, J.; Han, B. Pooling revisited: Your receptive field is suboptimal. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, New Orleans, LA, USA, 18–24 June 2022; IEEE: Piscataway, NJ, USA, 2022; pp. 549–558. [Google Scholar]
- Rezatofighi, H.; Tsoi, N.; Gwak, J.; Sadeghian, A.; Reid, I.; Savarese, S. Generalized intersection over union: A metric and a loss for bounding box regression. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Long Beach, CA, USA, 15–20 June 2019; IEEE: Piscataway, NJ, USA, 2019; pp. 658–666. [Google Scholar]
- Wang, J.; Xu, C.; Yang, W.; Yu, L. A normalized Gaussian Wasserstein distance for tiny object detection. arXiv 2021, arXiv:2110.13389. [Google Scholar]
- Lin, T.-Y.; Dollár, P.; Girshick, R.; He, K.; Hariharan, B.; Belongie, S. Feature pyramid networks for object detection. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, HI, USA, 21–26 June 2017; IEEE: Piscataway, NJ, USA, 2017; pp. 2117–2125. [Google Scholar]
- Neubeck, A.; Van Gool, L. Efficient non-maximum suppression. In Proceedings of the 18th international Conference on Pattern Recognition (ICPR’06), Hong Kong, China, 20–24 August 2006; pp. 850–855. [Google Scholar]
- Feng, C.; Zhong, Y.; Gao, Y.; Scott, M.R.; Huang, W. Tood: Task-aligned one-stage object detection. In Proceedings of the 2021 IEEE/CVF International Conference on Computer Vision (ICCV), Montreal, BC, Canada, 11–17 October 2021; IEEE: Piscataway, NJ, USA, 2021; pp. 3490–3499. [Google Scholar]
- Sun, Z.; Cao, S.; Yang, Y.; Kitani, K.M. Rethinking transformer-based set prediction for object detection. In Proceedings of the IEEE/CVF International Conference on Computer Vision, Montreal, BC, Canada, 11–17 October 2021; IEEE: Piscataway, NJ, USA, 2021; pp. 3611–3620. [Google Scholar]
- Dai, X.; Chen, Y.; Xiao, B.; Chen, D.; Liu, M.; Yuan, L.; Zhang, L. Dynamic head: Unifying object detection heads with attentions. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Montreal, BC, Canada, 11–17 October 2021; IEEE: Piscataway, NJ, USA, 2021; pp. 7373–7382. [Google Scholar]
- Zhu, X.; Lyu, S.; Wang, X.; Zhao, Q. TPH-YOLOv5: Improved YOLOv5 based on transformer prediction head for object detection on drone-captured scenarios. In Proceedings of the IEEE/CVF International Conference on Computer Vision, Montreal, BC, Canada, 11–17 October 2021; IEEE: Piscataway, NJ, USA, 2021; pp. 2778–2788. [Google Scholar]
- Meethal, A.; Granger, E.; Pedersoli, M. Cascaded zoom-in detector for high resolution aerial images. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Vancouver, BC, Canada, 17–24 June 2023; IEEE: Piscataway, NJ, USA, 2023; pp. 2045–2054. [Google Scholar]
- Sun, P.; Zhang, R.; Jiang, Y.; Kong, T.; Xu, C.; Zhan, W.; Tomizuka, M.; Li, L.; Yuan, Z.; Wang, C. Sparse r-cnn: End-to-end object detection with learnable proposals. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Montreal, BC, Canada, 11–17 October 2021; IEEE: Piscataway, NJ, USA, 2021; pp. 14454–14463. [Google Scholar]
- Li, J.; Liang, X.; Wei, Y.; Xu, T.; Feng, J.; Yan, S. Perceptual generative adversarial networks for small object detection. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, HI, USA, 21–26 June 2017; IEEE: Piscataway, NJ, USA, 2017; pp. 1222–1230. [Google Scholar]
- Bai, Y.; Zhang, Y.; Ding, M.; Ghanem, B. Sod-mtgan: Small object detection via multi-task generative adversarial network. In Proceedings of the European Conference on Computer Vision (ECCV), Munich, Germany, 8–14 September 2018; Springer: Cham, Switzerland, 2018; pp. 210–226. [Google Scholar]
- Dosovitskiy, A.; Beyer, L.; Kolesnikov, A.; Weissenborn, D.; Zhai, X.; Unterthiner, T.; Dehghani, M.; Minderer, M.; Heigold, G.; Gelly, S. An image is worth 16x16 words: Transformers for image recognition at scale. arXiv 2020, arXiv:2010.11929. [Google Scholar]
- Liu, Z.; Lin, Y.; Cao, Y.; Hu, H.; Wei, Y.; Zhang, Z.; Lin, S.; Guo, B. Swin transformer: Hierarchical vision transformer using shifted windows. In Proceedings of the IEEE/CVF International Conference on Computer Vision, Montreal, BC, Canada, 11–17 October 2021; IEEE: Piscataway, NJ, USA, 2021; pp. 10012–10022. [Google Scholar]
- Liu, S.; Li, F.; Zhang, H.; Yang, X.; Qi, X.; Su, H.; Zhu, J.; Zhang, L. Dab-detr: Dynamic anchor boxes are better queries for detr. arXiv 2022, arXiv:2201.12329. [Google Scholar]
- Hu, J.; Shen, L.; Sun, G. Squeeze-and-excitation networks. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, USA, 18–23 June 2018; IEEE: Piscataway, NJ, USA, 2018; pp. 7132–7141. [Google Scholar]
- Tolstikhin, I.O.; Houlsby, N.; Kolesnikov, A.; Beyer, L.; Zhai, X.; Unterthiner, T.; Yung, J.; Steiner, A.; Keysers, D.; Uszkoreit, J. Mlp-mixer: An all-mlp architecture for vision. Adv. Neural Inf. Process. Syst. 2021, 34, 24261–24272. [Google Scholar]
- Yu, J.; Jiang, Y.; Wang, Z.; Cao, Z.; Huang, T. Unitbox: An advanced object detection network. In Proceedings of the 24th ACM International Conference on Multimedia, Amsterdam, The Netherlands, 15–19 October 2016; Machinery: New York, NY, USA, 2016; pp. 516–520. [Google Scholar]
- Zheng, Z.; Wang, P.; Liu, W.; Li, J.; Ye, R.; Ren, D. Distance-IoU loss: Faster and better learning for bounding box regression. In Proceedings of the AAAI Conference on Artificial Intelligence, New York, NY, USA, 7–12 February 2020; pp. 12993–13000. [Google Scholar]
- Zhang, Y.-F.; Ren, W.; Zhang, Z.; Jia, Z.; Wang, L.; Tan, T. Focal and efficient IOU loss for accurate bounding box regression. Neurocomputing 2022, 506, 146–157. [Google Scholar] [CrossRef]
- Lin, T.-Y.; Goyal, P.; Girshick, R.; He, K.; Dollár, P. Focal loss for dense object detection. In Proceedings of the IEEE international Conference on Computer Vision; IEEE: Piscataway, NJ, USA, 2017; pp. 2980–2988. [Google Scholar]
- Jiang, B.; Luo, R.; Mao, J.; Xiao, T.; Jiang, Y. Acquisition of localization confidence for accurate object detection. In Proceedings of the European Conference on Computer Vision (ECCV), Munich, Germany, 8–14 September 2018; Springer: Cham, Switzerland, 2018; pp. 816–832. [Google Scholar]
- Du, D.; Zhu, P.; Wen, L.; Bian, X.; Lin, H.; Hu, Q.; Peng, T.; Zheng, J.; Wang, X.; Zhang, Y. VisDrone-DET2019: The vision meets drone object detection in image challenge results. In Proceedings of the IEEE/CVF International Conference on Computer Vision Workshops, Seoul, Republic of Korea, 27–28 October 2019; IEEE: Piscataway, NJ, USA, 2019. [Google Scholar]
- Wei, S.; Zeng, X.; Qu, Q.; Wang, M.; Su, H.; Shi, J. HRSID: A high-resolution SAR images dataset for ship detection and instance segmentation. IEEE Access 2020, 8, 120234–120254. [Google Scholar] [CrossRef]
- Everingham, M.; Van Gool, L.; Williams, C.K.; Winn, J.; Zisserman, A. The pascal visual object classes (voc) challenge. Int. J. Comput. Vis. 2010, 88, 303–338. [Google Scholar] [CrossRef]
- Lin, T.-Y.; Maire, M.; Belongie, S.; Hays, J.; Perona, P.; Ramanan, D.; Dollár, P.; Zitnick, C.L. Microsoft coco: Common objects in context. In Proceedings of the Computer Vision–ECCV 2014: 13th European Conference, Zurich, Switzerland, 6–12 September 2014; Proceedings, Part V 13. pp. 740–755. [Google Scholar]
- Xia, G.-S.; Bai, X.; Ding, J.; Zhu, Z.; Belongie, S.; Luo, J.; Datcu, M.; Pelillo, M.; Zhang, L. DOTA: A large-scale dataset for object detection in aerial images. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, USA, 18–23 June 2018; IEEE: Piscataway, NJ, USA, 2018; pp. 3974–3983. [Google Scholar]
- Loshchilov, I.; Hutter, F. Decoupled Weight Decay Regularization. arXiv 2017, arXiv:1711.05101. [Google Scholar]
- Lv, W.; Xu, S.; Zhao, Y.; Wang, G.; Wei, J.; Cui, C.; Du, Y.; Dang, Q.; Liu, Y. Detrs beat yolos on real-time object detection. arXiv 2023, arXiv:2304.08069. [Google Scholar]
- Yang, C.; Huang, Z.; Wang, N. Querydet: Cascaded sparse query for accelerating high-resolution small object detection. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, New Orleans, LA, USA, 18–24 June 2022; IEEE: Piscataway, NJ, USA, 2022; pp. 13668–13677. [Google Scholar]
- Xu, S.; Wang, X.; Lv, W.; Chang, Q.; Cui, C.; Deng, K.; Wang, G.; Dang, Q.; Wei, S.; Du, Y. PP-YOLOE: An evolved version of YOLO. arXiv 2022, arXiv:2203.16250. [Google Scholar]
- Zhang, X.; Izquierdo, E.; Chandramouli, K. Dense and small object detection in uav vision based on cascade network. In Proceedings of the IEEE/CVF International Conference on Computer Vision Workshops, Seoul, Republic of Korea, 27–28 October 2019; IEEE: Piscataway, NJ, USA, 2019. [Google Scholar]
Model | Epoch | AP | AP50 | APS | APM | APL |
---|---|---|---|---|---|---|
Deformable DETR [10] | 12 | 41.0 | 62.6 | 26.4 | 47.1 | 58.0 |
DAB−DETR [35] | 12 | 38.2 | 60.3 | 19.2 | 40.9 | 55.4 |
DN−DETR [13] | 12 | 43.4 | 61.9 | 24.8 | 46.8 | 59.4 |
DINO−4scale [17] | 12 | 49.1 | 66.6 | 32 | 52.3 | 63.2 |
DINO−4scale−SAC | 12 | 50.1 | 67.6 | 32.4 | 53.2 | 65.5 |
Settings | AP | AP50 | AP75 | APS | APM | APL |
---|---|---|---|---|---|---|
Baseline | 27.4 | 50.8 | 25.5 | 17.5 | 42.0 | 51.4 |
complete feature set | 28.6 | 52.3 | 26.9 | 21.0 | 38.9 | 45.4 |
low-level feature | 30.9 | 53.9 | 30.9 | 22.9 | 41.7 | 52.6 |
IOU_Ratio () | AP | AP50 | AP75 | APS | APM | APL |
---|---|---|---|---|---|---|
0 | 27.6 | 50.8 | 26.0 | 17.6 | 42.3 | 50.6 |
0.3 | 29.3 | 53.5 | 27.5 | 20.2 | 41.2 | 51.9 |
0.5 | 28.4 | 52.4 | 26.9 | 19.7 | 40.2 | 49.4 |
1 | 28.3 | 52.3 | 26.2 | 19.8 | 40.3 | 48.3 |
SAC | RSP | NWD | AP | APS | APM | APL |
---|---|---|---|---|---|---|
27.4 | 17.5 | 42.0 | 51.4 | |||
√ | 28.2 | 18.3 | 41.2 | 53.9 | ||
√ | 31.8 | 23.8 | 41.3 | 50.7 | ||
√ | 29.5 | 20.5 | 41.5 | 50.2 | ||
√ | √ | √ | 33.2 | 24.8 | 41.8 | 50.8 |
SAC | RSP | NWD | AP | APS | APM | APL |
---|---|---|---|---|---|---|
57.7 | 58.1 | 60.6 | 31.8 | |||
√ | 58.3 | 58.6 | 60.8 | 31.9 | ||
√ | 63.5 | 64.6 | 63.4 | 25.3 | ||
√ | 59.4 | 59.9 | 63.9 | 24.4 | ||
√ | √ | √ | 65.4 | 66.6 | 64.0 | 37.6 |
Model | Backbone | AP | AP50 | AP75 | APS | APM | APL | FPS |
---|---|---|---|---|---|---|---|---|
FCOS [15] | ResNet–50 | 29.5 | 50.4 | 29.9 | 21.3 | 40.5 | 37.3 | 18 |
RetinaNet [41] | ResNet–50 | 26.2 | 44.9 | 27.1 | 18.6 | 37.5 | 43.8 | 14 |
QueryDet [50] | ResNet–50 | 28.3 | 48.1 | 28.8 | 20.2 | 37.2 | 46.1 | 16 |
CascadeNet [52] | ResNet–50 | 28.8 | 47.1 | 29.3 | 18.9 | 38.1 | 47.6 | 22 |
FRCNN+FPN [4] | ResNet–50 | 21.9 | 37.6 | 22.4 | 13.9 | 32.9 | 47.5 | 21 |
YOLOv5 | CSP–Darknet53 | 26.0 | 42.7 | 27.1 | 15.6 | 42.1 | 52.2 | 43 |
PP–YOLOE–l [51] | CSPRepResNet | 29.2 | 47.3 | 30.1 | 18.4 | 44.0 | 63.3 | 94 |
PP–YOLOE+_SOD–l | CSPRepResNet | 31.9 | 52.1 | 32.6 | 21.7 | 45.1 | 60.8 | 90 |
RT–DETR–X [49] | ResNet–50 | 31.0 | 52.0 | 30.9 | 21.2 | 42.9 | 61.8 | 108 |
OURS | ResNet–50 | 33.2 | 56.4 | 33.0 | 24.8 | 43.8 | 50.8 | 15 |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Xiao, X.; Xue, X.; Zhao, Z.; Fan, Y. A Recursive Prediction-Based Feature Enhancement for Small Object Detection. Sensors 2024, 24, 3856. https://doi.org/10.3390/s24123856
Xiao X, Xue X, Zhao Z, Fan Y. A Recursive Prediction-Based Feature Enhancement for Small Object Detection. Sensors. 2024; 24(12):3856. https://doi.org/10.3390/s24123856
Chicago/Turabian StyleXiao, Xiang, Xiaorong Xue, Zhiyuan Zhao, and Yisheng Fan. 2024. "A Recursive Prediction-Based Feature Enhancement for Small Object Detection" Sensors 24, no. 12: 3856. https://doi.org/10.3390/s24123856