![[2602.13378] LAF-YOLOv10 with Partial Convolution Backbone, Attention-Guided Feature Pyramid, Auxiliary P2 Head, and Wise-IoU Loss for Small Object Detection in Drone Aerial Imagery](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.13378] LAF-YOLOv10 with Partial Convolution Backbone, Attention-Guided Feature Pyramid, Auxiliary P2 Head, and Wise-IoU Loss for Small Object Detection in Drone Aerial Imagery
Summary
The paper presents LAF-YOLOv10, an advanced model for small object detection in drone imagery, integrating techniques like Partial Convolution and Attention-Guided Feature Pyramid to enhance performance in challenging conditions.
Why It Matters
As drone technology becomes increasingly prevalent in surveillance and monitoring, improving small object detection is crucial. This research addresses specific challenges faced in aerial imagery, potentially enhancing applications in disaster response and traffic monitoring.
Key Takeaways
- LAF-YOLOv10 improves small object detection in drone imagery.
- Integrates Partial Convolution and Attention-Guided Feature Pyramid techniques.
- Achieves higher mAP scores compared to previous models, indicating better performance.
- Demonstrates practical viability for embedded UAV deployment with real-time processing.
- Addresses unique challenges of UAV imagery, such as occlusion and small target size.
Computer Science > Computer Vision and Pattern Recognition arXiv:2602.13378 (cs) [Submitted on 13 Feb 2026] Title:LAF-YOLOv10 with Partial Convolution Backbone, Attention-Guided Feature Pyramid, Auxiliary P2 Head, and Wise-IoU Loss for Small Object Detection in Drone Aerial Imagery Authors:Sohail Ali Farooqui, Zuhair Ahmed Khan Taha, Mohammed Mudassir Uddin, Shahnawaz Alam View a PDF of the paper titled LAF-YOLOv10 with Partial Convolution Backbone, Attention-Guided Feature Pyramid, Auxiliary P2 Head, and Wise-IoU Loss for Small Object Detection in Drone Aerial Imagery, by Sohail Ali Farooqui and 3 other authors View PDF HTML (experimental) Abstract:Unmanned aerial vehicles serve as primary sensing platforms for surveillance, traffic monitoring, and disaster response, making aerial object detection a central problem in applied computer vision. Current detectors struggle with UAV-specific challenges: targets spanning only a few pixels, cluttered backgrounds, heavy occlusion, and strict onboard computational budgets. This study introduces LAF-YOLOv10, built on YOLOv10n, integrating four complementary techniques to improve small-object detection in drone imagery. A Partial Convolution C2f (PC-C2f) module restricts spatial convolution to one quarter of backbone channels, reducing redundant computation while preserving discriminative capacity. An Attention-Guided Feature Pyramid Network (AG-FPN) inserts Squeeze-and-Excitation channel gates before multi-scale fusion and replaces...