Research map / Detection and segmentation / CNN features
Recurrent Vision Transformers for Object Detection With Event Cameras
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 RVT-B instead of ASTMNet | Event-camera object detection · Gen1 test | mAP: 46.7 -> 47.2 | Helped |
| Use RVT-B instead of ASTMNet | Event-camera object detection · 1 Mpx test | mAP: 48.3 -> 47.4 | Hurt |
| Replace ConvNeXt blocks with multi-axis attention (ablation) | RVT spatial aggregation · Gen1 validation | mAP: 45.5 -> 47.6 | Helped |
| Replace Swin blocks with multi-axis attention (ablation) | RVT spatial aggregation · 1 Mpx validation | mAP: 44.4 -> 46.0 | Helped |
| Use non-overlapping instead of overlapping convolution kernels (ablation) | RVT downsampling · Gen1 validation | mAP: 47.6 -> 46.1 | Hurt |
| Use a 3×3 LSTM kernel instead of a 1×1 LSTM (ablation) | RVT temporal aggregation · Gen1 validation | mAP: 47.6 -> 46.5 | Hurt |
| Use a 3×3 depthwise-separable Conv-LSTM instead of a 1×1 LSTM (ablation) | RVT temporal aggregation · Gen1 validation | mAP: 47.6 -> 46.3 | Hurt |
| Add recurrence to S4, then S3, S2, and S1 (ablation) | RVT backbone · Gen1 validation | mAP: 32.0 -> 39.8 -> 44.2 -> 46.9 -> 47.6 | Helped |
| Use horizontal flipping, zoom-in, and zoom-out instead of no augmentation (ablation) | RVT training · Gen1 validation | mAP: 38.1 -> 47.6 | Helped |
| Use horizontal flipping only (ablation) | RVT training · Gen1 validation | mAP: 38.1 -> 41.6 | Helped |
| Use zoom-in only (ablation) | RVT training · Gen1 validation | mAP: 38.1 -> 45.8 | Helped |
| Use zoom-out only (ablation) | RVT training · Gen1 validation | mAP: 38.1 -> 44.1 | Helped |
Closest papers by meaning
- Scene Adaptive Sparse Transformer for Event-based Object Detection (CVPR 2024 · similarity 0.88)
- FlexEvent: Towards Flexible Event-Frame Object Detection at Varying Operational Frequencies (NeurIPS 2025 · similarity 0.87)
- LEOD: Label-Efficient Object Detection for Event Cameras (CVPR 2024 · similarity 0.87)
- EGSST: Event-based Graph Spatiotemporal Sensitive Transformer for Object Detection (NeurIPS 2024 · similarity 0.87)
- Event Stream-based Visual Object Tracking: A High-Resolution Benchmark Dataset and A Novel Baseline (CVPR 2024 · similarity 0.87)
- Hybrid Spiking Vision Transformer for Object Detection with Event Cameras (ICML 2025 · similarity 0.86)
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.