Research map / Detection and segmentation / CNN features
Adaptive Sparse Convolutional Networks With Global Context Enhancement for Faster Object Detection on Drone Images
Bowei Du, Yecheng Huang, Jiaxin Chen, Di Huang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| replace the GFL V1 detection head with CEASC | GFL V1, ResNet18 · VisDrone | GFLOPs: 524.95 -> 150.18 | Improved |
| replace the GFL V1 detection head with CEASC | GFL V1, ResNet18 · VisDrone | mAP: 28.4 -> 28.7 | Improved |
| replace the detection head with CEASC | GFL V1 · UAVDT | mAP: 16.9 -> 17.1 | Improved |
| add CESC to the dense baseline (ablation) | GFL V1 · VisDrone | GFLOPs: 524.95 -> 158.23 | Improved |
| add adaptive multi-layer masking to CESC (ablation) | GFL V1 with CESC · VisDrone | mAP: 28.6 -> 28.7 | Improved |
| add CE-GN to sparse convolution with a residual structure (ablation) | GFL V1 with sparse convolution · VisDrone | mAP: 26.1 -> 28.5 | Improved |
| add normalization loss to sparse convolution, residual, and CE-GN (ablation) | GFL V1 with CESC · VisDrone | mAP: 28.5 -> 28.7 | Improved |
| replace CE-GN with GroupNorm (ablation) | GFL V1 with sparse convolution · VisDrone | mAP: 28.7 -> 28.0 | Hurt |
| replace CE-GN with BatchNorm (ablation) | GFL V1 with sparse convolution · VisDrone | mAP: 28.7 -> 26.1 | Hurt |
| replace CE-GN with InstanceNorm (ablation) | GFL V1 with sparse convolution · VisDrone | mAP: 28.7 -> 27.9 | Hurt |
| replace point-wise global-context encoding with 3×3 convolution (ablation) | GFL V1 with CESC · VisDrone | mAP: 28.7 -> 28.5 | Hurt |
| replace point-wise global-context encoding with GhostModule (ablation) | GFL V1 with CESC · VisDrone | mAP: 28.7 -> 28.3 | Hurt |
Closest papers by meaning
- RemDet: Rethinking Efficient Model Design for UAV Object Detection (AAAI 2025 · similarity 0.88)
- DroneFINE: Domain-Aware Parameter-Efficient Fine-Tuning of Vision-Language Detectors for Drone Images (ECCV 2026 · similarity 0.87)
- Delving Into Robust Object Detection From Unmanned Aerial Vehicles: A Deep Nuisance Disentanglement Approach (ICCV 2019 · similarity 0.86)
- PointSR: Self-Regularized Point Supervision for Drone-View Object Detection (CVPR 2025 · similarity 0.85)
- SparseAlign: a Fully Sparse Framework for Cooperative Object Detection (CVPR 2025 · similarity 0.85)
- Multiview Aerial Visual RECognition (MAVREC): Can Multi-view Improve Aerial Visual Perception? (CVPR 2024 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
