Combining Transformers and CNNs for Efficient Object Detection in High-Resolution Satellite Imagery
Not on the map. Why?
Abstract (arXiv)
We present GLOD, a transformer-first architecture for object detection in high-resolution satellite imagery. GLOD replaces CNN backbones with a Swin Transformer for end-to-end feature extraction, combined with novel UpConvMixer blocks for robust upsampling and Fusion Blocks for multi-scale feature integration. Our approach achieves 32.95% on xView, outperforming SOTA methods by 11.46%. Key innovations include asymmetric fusion with CBAM attention and a multi-path head design capturing objects across scales. The architecture is optimized for satellite imagery challenges, leveraging spatial priors while maintaining computational efficiency.
Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Builds on 35 of its 45 references that are in the map. First posted on arXiv 15 Jul 2025.
Builds on
35 of 45 references.
References · 35
- 2017Attention Is All You Needcited by 60,361
- 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 30,186
- 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,713
- 2014Microsoft COCO: Common Objects in Contextcited by 17,608
- 2021Swin Transformer: Hierarchical Vision Transformer using Shifted Windowscited by 12,139
- 2015Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networkscited by 9,639
- 2020End-to-End Object Detection with Transformerscited by 7,414
- 2021Masked Autoencoders Are Scalable Vision Learnerscited by 7,403
- 2021Emerging Properties in Self-Supervised Vision Transformerscited by 6,968
- 2017mixup: Beyond Empirical Risk Minimizationcited by 5,183
- 2016Feature Pyramid Networks for Object Detectioncited by 5,036
- 2015You Only Look Once: Unified, Real-Time Object Detectioncited by 4,852
23 more
- 2020Training data-efficient image transformers & distillation through attentioncited by 4,586
- 2015SSD: Single Shot MultiBox Detectorcited by 3,130
- 2015Fast R-CNNcited by 2,933
- 2018CBAM: Convolutional Block Attention Modulecited by 2,477
- 2013Rich feature hierarchies for accurate object detection and semantic segmentationcited by 2,162
- 2017Rethinking Atrous Convolution for Semantic Image Segmentationcited by 2,161
- 2019Deep High-Resolution Representation Learning for Visual Recognitioncited by 1,420
- 2019Generalized Intersection over Union: A Metric and A Loss for Bounding Box Regressioncited by 1,368
- 2021Swin Transformer V2: Scaling Up Capacity and Resolutioncited by 1,228
- 2016Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Networkcited by 1,204
- 2019Objects as Pointscited by 1,175
- 2015Deep Learning and the Information Bottleneck Principlecited by 802
- 2021CoAtNet: Marrying Convolution and Attention for All Data Sizescited by 569
- 2019Distance-IoU Loss: Faster and Better Learning for Bounding Box Regressioncited by 459
- 2022DiffusionDet: Diffusion Model for Object Detectioncited by 354
- 2021ResMLP: Feedforward networks for image classification with data-efficient trainingcited by 339
- 2020DetectoRS: Detecting Objects with Recursive Feature Pyramid and Switchable Atrous Convolutioncited by 222
- 2015Training Very Deep Networkscited by 173
- 2022Slicing Aided Hyper Inference and Fine-tuning for Small Object Detectioncited by 105
- 2022Detecting tiny objects in aerial images: A normalized Wasserstein distance and a new benchmarkcited by 56
- 2024DQ-DETR: DETR with Dynamic Query for Tiny Object Detectioncited by 28
- 2024A DeNoising FPN With Transformer R-CNN for Tiny Object Detectioncited by 12
- 2019Reduced Focal Loss: 1st Place Solution to xView object detection in Satellite Imagerycited by 8
Built on by
Citing papers per year: 2025: 1.
1 paper.