A Global-Local Self-Adaptive Network for Drone-View Object Detection

Deng, Sutao; Li, Shuai; Xie, Ke; Song, Wenfeng; Xiao, Liao; Hao, Aimin; Qin, Hong

doi:10.1109/tip.2020.3045636

Cited by 116 publications

(59 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Compared to DREN [44] and ResNet101, the AP is increased by 4.5% with ResNet-50 backbone and 2.4% with ResNet-101 backbone on UAVDT dataset. Compared with GLSAN [8], which uses extra super-resolution network for enlarging regions, our method outperforms it by 5.49% of AP on VisDrone2019. The performance of the coarse-to-fine methods is limited by the initial detection, which would cause the bias for region generation (i.e., small objects are easily missed in coarse-level detection).…”

Section: Comparison With State-of-the-art Modelsmentioning

confidence: 96%

“…On UAVDT dataset, we achieve 22.4% of AP and 38.6% of AP 50 with Cascade R-CNN as the detector. Compared to GLSAN [8] with Cascade R-CNN, the AP is increased by 3.4% and AP 50 is increased by 8.1%. On Visdrone dataset, the AdaZoom achieves 40.33% and AP 50 of 66.94%, outperforming the SOTA performance by a large margin.…”

Section: Comparison With State-of-the-art Modelsmentioning

confidence: 98%

“…Comparison with coarse-to-fine detection methods. For such methods, the object distribution is estimated based on a coarse-level preview detection, such as DREN [44], CRENet [40] and GLSAN [8]. Compared to CRENet with backbone of Hourglass-104, Our AdaZoom with Faster RCNN achieves 2.49% higher in AP on VisDrone2019 with ResNet-50 backbone.…”

Section: Comparison With State-of-the-art Modelsmentioning

confidence: 99%

See 2 more Smart Citations

AdaZoom: Adaptive Zoom Network for Multi-Scale Object Detection in Large Scenes

Xu¹,

Li²,

Wang³

2021

Preprint

View full text Add to dashboard Cite

Detection in large-scale scenes is a challenging problem due to small objects and extreme scale variation. It is essential to focus on the image regions of small objects. In this paper, we propose a novel Adaptive Zoom (AdaZoom) network as a selective magnifier with flexible shape and focal length to adaptively zoom the focus regions for object detection. Based on policy gradient, we construct a reinforcement learning framework for focus region generation, with the reward formulated by object distributions. The scales and aspect ratios of the generated regions are adaptive to the scales and distribution of objects inside. We apply variable magnification according to the scale of the region for adaptive multi-scale detection. We further propose collaborative training to complementarily promote the performance of AdaZoom and detection network. To validate the effectiveness, we conduct extensive experiments on VisDrone2019, UAVDT and DOTA datasets. The experiments show AdaZoom brings consistent and significant improvement over different detection networks, achieving state-of-the-art performance on these datasets, especially outperforming the existing methods by AP of 4.64% on Vis-Drone2019.

show abstract

Section: Comparison With State-of-the-art Modelsmentioning

confidence: 96%

Section: Comparison With State-of-the-art Modelsmentioning

confidence: 98%

Section: Comparison With State-of-the-art Modelsmentioning

confidence: 99%

See 1 more Smart Citation

AdaZoom: Adaptive Zoom Network for Multi-Scale Object Detection in Large Scenes

Xu¹,

Li²,

Wang³

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Reference [ 36 ] proposed a depthwise separable attention-guided network (DAGN), which integrated the feature series with a concentration block to make sure that the model is able to brilliantly differentiate significant and trivial features. Reference [ 37 ] integrated the overall and partial fusion strategy with a progressive network with varying scales to fullfill detection in a more accurate manner. In [ 7 ], an anchor-free method was introduced.…”

Section: Related Workmentioning

confidence: 99%

Lightweight Detection Network Based on Sub-Pixel Convolution and Objectness-Aware Structure for UAV Images

Jiang

et al. 2021

Sensors

View full text Add to dashboard Cite

Unmanned Aerial Vehicles (UAVs) can serve as an ideal mobile platform in various situations. Real-time object detection with on-board apparatus provides drones with increased flexibility as well as a higher intelligence level. In order to achieve good detection results in UAV images with complex ground scenes, small object size and high object density, most of the previous work introduced models with higher computational burdens, making deployment on mobile platforms more difficult.This paper puts forward a lightweight object detection framework. Besides being anchor-free, the framework is based on a lightweight backbone and a simultaneous up-sampling and detection module to form a more efficient detection architecture. Meanwhile, we add an objectness branch to assist the multi-class center point prediction, which notably improves the detection accuracy and only takes up very little computing resources. The results of the experiment indicate that the computational cost of this paper is 92.78% lower than the CenterNet with ResNet18 backbone, and the mAP is 2.8 points higher on the Visdrone-2018-VID dataset. A frame rate of about 220 FPS is achieved. Additionally, we perform ablation experiments to check on the validity of each part, and the method we propose is compared with other representative lightweight object detection methods on UAV image datasets.

show abstract

“…Yang et al 2019), DMNet(Li et al 2020a), GLSAN(Deng et al 2021), and DREN (Zhang et al 2019). Results on VisDrone.…”

mentioning

confidence: 99%

UFPMP-Det: Toward Accurate and Efficient Object Detection on Drone Imagery

Huang¹,

Chen²,

Huang³

2021

Preprint

View full text Add to dashboard Cite

This paper proposes a novel approach to object detection on drone imagery, namely Multi-Proxy Detection Network with Unified Foreground Packing (UFPMP-Det). To deal with the numerous instances of very small scales, different from the common solution that divides the high-resolution input image into quite a number of chips with low foreground ratios to perform detection on them each, the Unified Foreground Packing (UFP) module is designed, where the sub-regions given by a coarse detector are initially merged through clustering to suppress background and the resulting ones are subsequently packed into a mosaic for a single inference, thus significantly reducing overall time cost. Furthermore, to address the more serious confusion between inter-class similarities and intraclass variations of instances, which deteriorates detection performance but is rarely discussed, the Multi-Proxy Detection Network (MP-Det) is presented to model object distributions in a fine-grained manner by employing multiple proxy learning, and the proxies are enforced to be diverse by minimizing a Bag-of-Instance-Words (BoIW) guided optimal transport loss. By such means, UFPMP-Det largely promotes both the detection accuracy and efficiency. Extensive experiments are carried out on the widely used VisDrone and UAVDT datasets, and UFPMP-Det reports new state-of-the-art scores at a much higher speed, highlighting its advantages.

show abstract

A Global-Local Self-Adaptive Network for Drone-View Object Detection

Cited by 116 publications

References 36 publications

AdaZoom: Adaptive Zoom Network for Multi-Scale Object Detection in Large Scenes

AdaZoom: Adaptive Zoom Network for Multi-Scale Object Detection in Large Scenes

Lightweight Detection Network Based on Sub-Pixel Convolution and Objectness-Aware Structure for UAV Images

UFPMP-Det: Toward Accurate and Efficient Object Detection on Drone Imagery

Contact Info

Product

Resources

About