Aug 2026· PLoS ONE· Vol 21, pp. e0352244· 0 citations· 36 references
Medicine
TL;DR
Improved PF-DETR improves the model’s robustness and accuracy in detecting multi-scale and small targets in complex and cluttered scenes, resulting in a favorable balance between detection performance and model efficiency.
Abstract
To tackle the challenge of detecting small targets in UAV imagery, this paper proposes PF-DETR, an enhanced object detection model based on RT-DETR, designed to improve detection accuracy in complex scenes. The improvements are primarily reflected in the following aspects. First, a P2 detection head is added to extend the feature pyramid to finer scales, thereby enhancing the ability of shallow features to detect small targets. Second, a Pyramidal Hierarchical Frequency-Domain Fusion (PHF) module is introduced. By combining wavelet pooling with high- and low-frequency attention, the module effectively extracts and fuses multi-scale features, reduces information loss, and improves detection accuracy for small targets. Finally, the backbone network is restructured through the design of a lightweight BasicBlock_FasterNet_Rep module, which integrates FasterNet and RepVGG-style re-parameterization. This restructuring significantly reduces model complexity and parameters while strengthening multi-scale feature extraction. Experimental results on the VisDrone2019 dataset show that the improved PF-DETR achieves a notable performance boost: compared to the original model, mAP@0.5 increases by 5.4%, while the number of parameters is reduced by 25.0%. The computational cost increases by about 35.0%, but this comes with higher accuracy, resulting in a favorable balance between detection performance and model efficiency. Overall, these improvements enhance the model’s robustness and accuracy in detecting multi-scale and small targets in complex and cluttered scenes.
To address the challenges of UAV aerial imagery, including the prevalence of small objects, complex background interference, and difficulty in feature extraction that lead to high missed detection rates and compromise detection accuracy in existing RT-DETR algorithms, this paper proposes an improved small-object-oriented detector named MSFE-DETR (Multi-Scale Feature Enhancement DETR). A CMFE (CSP-MultiScale Feature Enhancement) module is integrated into the shallow backbone layers to enhance feature representation of small objects and alleviate feature loss caused by scale and background complexity. In deeper layers of backbone, the C2f module is employed to preserve fine-grained details and improve target–background discrimination, while multi-scale feature fusion further prevents small object information degradation. In addition, Deformable Attention (DAttention) is incorporated to adaptively focus on small target regions, retaining spatial positional information and suppressing background noise. The head integrates MPCA and FSA modules, where MPCA progressively fuses adjacent-scale features to complementarily enhance small object representations and suppress background interference, and FSA further improves detail enhancement and robustness. Moreover, an Inner-SIoU loss is proposed by combining Inner-IoU with SIoU, improving localization accuracy, convergence speed, and robustness in complex scenes. Experimental results on the VisDrone 2019 dataset show that MSFE-DETR outperforms RT-DETR-r18 by 1.9% in Precision, 2.1% in Recall and 2.4% in mAP@0.5, while real-time inference is maintained at 68.7 FPS. On the infrared HIT-UAV and satellite SIMD datasets, mAP@0.5 improves by 5.8% and 2.3% respectively over the baseline.
A gated feature fusion net is designed to reduce model parameters and a small object detection pyramid is implemented to preserve high-resolution details and integrate a mini-kernel that utilizes a lossless down-sampling module for deep feature optimization.
Jun-Qiu Wang, Xiang-Yang Lu, Yanda Wang et al.· 0 citations
Small-object detection in unmanned aerial vehicle (UAV) imagery remains challenging because target objects often occupy only a few pixels, exhibit weak feature responses, and are easily obscured by complex backgrounds. These aspects significantly limit the effectiveness of end-to-end detection systems. To overcome these limitations and enhance the detection accuracy in challenging UAV settings, this paper proposes MPC-DETR, a Multi-scale Patch Context Transformer that is based on RT-DETR. To begin with, a Local-Global Attention Fusion Module (LGAF) is proposed to capture fine-grained local features and long-range semantic relations of small objects. LGAF enhances feature representation through a lightweight multi-branch synergistic attention mechanism while introducing limited computational overhead. Second, a Dilated Context-Aware Feature Interaction Module (DCFI) is proposed to enhance the discriminative capability of high-level features in cluttered backgrounds and densely populated small-object scenes. DCFI allows more efficient feature aggregation and contextual comprehension through multi-scale contextual modeling and scale-adaptive feature interaction. Third, a Patch-Guided Multi-scale Feature Fusion Module (PGMFF) is developed to create a patch-guided contextual fusion approach that combines shallow, high-resolution features with deeper semantic information. This process improves the maintenance and representation of fine object information and minimizes information loss in feature propagation. The experimental results on the VisDrone2019 dataset show that MPC-DETR has an mAP50 and mAP50–95 of 52.5% and 33.3%, respectively, which are 4.6 and 4.0 percentage points higher than the baseline model. Further analyses of the UAVDT and HIT-UAV datasets also support the high generalization potential of the suggested method to various UAV-based small-object detection problems. In general, the findings suggest that MPC-DETR provides precise, strong, and efficient small-object detection in complicated UAV images.
Accurate detection of small-scale targets in UAV aerial imagery remains challenging due to severe scale variation, background interference, and feature degradation caused by repeated downsampling. To address these issues, this study proposes YOLO-SWIFT, a wavelet-integrated feature transformation network designed for small-object detection. First, a Position-Aware Coordinate Downsampling module is developed to preserve critical spatial information during feature compression through coordinate attention and skip connections. Second, a High-Resolution Feature Aggregation Network is introduced to establish an additional high-resolution detection branch for enhanced cross-scale feature fusion. Third, a Wavelet Bottleneck Enhancement module incorporating multi-level wavelet decomposition and a High-Frequency Retention pathway is designed to improve fine-detail representation while expanding the effective receptive field. Finally, an Adaptive Scale-aware Regression IoU loss function is proposed to dynamically balance localization and shape-consistency constraints for small targets. Experimental evaluation on the VisDrone2019 benchmark demonstrates that YOLO-SWIFT achieves 38.7% mAP50 and 22.5% mAP50:95. The proposed framework provides an effective solution for intelligent aerial sensing and offers potential applications in electromagnetic imaging, remote sensing interpretation, and autonomous surveillance systems.
Y.-C. Li, K. Shao· Advanced Electromagnetics· 0 citations
HD-YOLO improves small-object detection with a compact parameter footprint, while direct hardware benchmarks remain necessary to establish deployment efficiency.
Maosheng Sun, Jing Ding, Yang Zhang et al.· Applied Sciences· 0 citations
Low-altitude UAV aerial imagery often has complex backgrounds with densely distributed small objects, posing challenges to accurate small-object detection. To address these problems, we propose a spatial–frequency response aware synergistic network for small-object detection in low-altitude UAV aerial imagery. A Frequency-Response-Aware Enhancement Module (FRAEM) is designed to effectively extract discriminative features. The module employs a deterministic stage-aware filtering strategy: Scharr-based edge-sensitive filtering is used in the shallow stage, whereas Gaussian smoothing is used in deeper stages, enabling complementary enhancement of hierarchical representations. A Detail Feature Fusion module (DFFusion) is then developed to improve the efficiency of multi-scale feature fusion. The existing Content-Aware Reassembly of Features (CARAFE) operator is employed for content-aware upsampling and feature alignment, after which DFFusion uses learnable scalar weighting to integrate high-resolution detail information with low-resolution contextual information. A Lightweight Adaptive Decoupled Head (LADH) is also designed to reduce complexity. LADH asymmetrically allocates computational capacity across the prediction tasks: the confidence branch retains stronger spatial processing, whereas the classification and regression branches use lightweight projections; depthwise separable convolution serves as an efficiency-oriented implementation choice. Experiments on the VisDrone2019 and DOTA-v2.0 datasets demonstrate that the proposed method can achieve balance between detection performance and model complexity over mainstream detection methods. Ablation experiments also prove the effectiveness of the proposed components.
