Research map / Vision-language and multimodal / Human and embodied modeling
LSKNet: A Foundation Lightweight Backbone for Remote Sensing
Yuxuan Li, Xiang Li, Yimain Dai, Qibin Hou, Li Liu, Yongxiang Liu, Ming‐Ming Cheng, Jian Yu Yang
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 channel selection with spatial selection (ablation) | LSKNet · DOTA-v1.0 | mAP: 80.57 -> 81.31 | Improved |
| add channel selection on top of spatial selection (ablation) | LSKNet · DOTA-v1.0 | mAP: 81.31 -> 80.82 | Hurt |
| add spatial selection to series large kernels (RF 23) (ablation) | LSKNet · DOTA-v1.0 | mAP: 80.94 -> 81.31 | Improved |
| use series instead of parallel kernel flow (ablation) | LSKNet vs SKNet-style · DOTA-v1.0 | mAP: 80.19 -> 80.57 | Improved |
| decompose 29x29 kernel into two depth-wise kernels (ablation) | LSKNet · DOTA-v1.0 | mAP: 80.66 -> 80.91 | Improved |
| decompose into three kernels instead of two (ablation) | LSKNet · DOTA-v1.0 | mAP: 80.91 -> 80.77 | Hurt |
| enlarge receptive field from 23 to 39 (ablation) | LSKNet · DOTA-v1.0 | mAP: 80.94 -> 80.84 | Hurt |
| add average pooling to max pooling in spatial selection (ablation) | LSKNet · DOTA-v1.0 | mAP: 81.23 -> 81.31 | Improved |
| replace ResNet-18 with LSKNet-T backbone | Oriented RCNN · DOTA-v1.0 | mAP: 79.27 -> 81.31 | Improved |
| replace ResNet-18 with LSKNet-T backbone | R3Det · DOTA-v1.0 | mAP: 74.16 -> 78.39 | Improved |
| replace ResNet-18 with LSKNet-T backbone | S2A-Net · DOTA-v1.0 | mAP: 76.82 -> 80.15 | Improved |
| replace ResNet-18 with LSKNet-T backbone | RoI Transformer · DOTA-v1.0 | mAP: 78.32 -> 80.89 | Improved |
Closest papers by meaning
- Large Selective Kernel Network for Remote Sensing Object Detection (ICCV 2023 · similarity 0.95)
- Poly Kernel Inception Network for Remote Sensing Detection (CVPR 2024 · similarity 0.86)
- Locate Anything on Earth: Advancing Open-Vocabulary Object Detection for Remote Sensing Community (AAAI 2025 · similarity 0.85)
- InstructSAM: A Training-free Framework for Instruction-Oriented Remote Sensing Object Recognition (NeurIPS 2025 · similarity 0.85)
- SkyScript: A Large and Semantically Diverse Vision-Language Dataset for Remote Sensing (AAAI 2024 · similarity 0.85)
- Towards Open-Vocabulary Remote Sensing Image Semantic Segmentation (AAAI 2025 · similarity 0.85)
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.
