Research map / Video / Video segmentation and tracking
LSTFE-Net:Long Short-Term Feature Enhancement Network for Video Small Object Detection
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Use LSTFE-Net to combine selected long-term context with aligned short-term features | Video object detection with ResNet101 · ImageNet-VID | mAP: 82.9 -> 83.4 | Helped |
| Use LSTFE-Net instead of TransVOD | Video small object detection with ResNet50 · FL-Drones | AP: 36.9 -> 37.8 | Helped |
| Use LSTFE-Net instead of MEGA | Video small object detection with ResNet101 · FL-Drones | AP: 42.4 -> 46.8 | Helped |
| Add one cascaded deformable convolution block for feature alignment (ablation) | LSTFE-Net with ResNet101 · FL-Drones | AP: 42.1 -> 45.4 | Helped |
| Increase alignment blocks from one to two (ablation) | LSTFE-Net with ResNet101 · FL-Drones | AP: 45.4 -> 39.4 | Hurt |
| Increase alignment blocks from one to three (ablation) | LSTFE-Net with ResNet101 · FL-Drones | AP: 45.4 -> 39.3 | Hurt |
| Select the long-term frame with lowest similarity to the current frame (ablation) | LSTFE-Net with ResNet101, 1 sampled long-term frame · FL-Drones | AP: 29.5 -> 29.9 | Helped |
| Select long-term frames by lowest similarity (ablation) | LSTFE-Net with ResNet101, 2 sampled long-term frames · FL-Drones | AP: 35.1 -> 35.9 | Helped |
| Select long-term frames by lowest similarity (ablation) | LSTFE-Net with ResNet101, 3 sampled long-term frames · FL-Drones | AP: 33.9 -> 34.8 | Helped |
| Select long-term frames by lowest similarity (ablation) | LSTFE-Net with ResNet101, 4 sampled long-term frames · FL-Drones | AP: 29.6 -> 30.7 | Helped |
| Replace concatenation with attention-based long short-term feature aggregation (ablation) | LSTFE-Net with ResNet101 · FL-Drones | AP: 39.9 -> 42.1 | Helped |
| Add location coding to short-term feature aggregation (ablation) | LSTFE-Net with ResNet101 · FL-Drones | AP: 41.2 -> 42.1 | Helped |
Closest papers by meaning
- Mobile Video Object Detection With Temporally-Aware Feature Maps (CVPR 2018 · similarity 0.87)
- Training-Time-Friendly Network for Real-Time Object Detection (AAAI 2020 · similarity 0.86)
- Video Object Detection with Locally-Weighted Deformable Neighbors (AAAI 2019 · similarity 0.86)
- Temporal Object-Aware Vision Transformer for Few-Shot Video Object Detection (AAAI 2026 · similarity 0.86)
- FSOD-VFM: Few-Shot Object Detection with Vision Foundation Models and Graph Diffusion (ICLR 2026 · similarity 0.86)
- Context Enhanced Transformer for Single Image Object Detection in Video Data (AAAI 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.