Research map / Detection and segmentation / CNN features
SODFormer: Streaming Object Detection With Transformer Using Events and Frames
Dianze Li, Yonghong Tian, Jianing Li
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 |
|---|---|---|---|
| add temporal transformer (ablation) | Deformable DETR RGB baseline · PKU-DAVIS-SOD test (all scenarios) | mAP50: 0.461 -> 0.489 | Improved |
| add DVS events with averaging fusion (ablation) | temporal transformer on frames · PKU-DAVIS-SOD test (all scenarios) | mAP50: 0.489 -> 0.494 | Improved |
| replace averaging with asynchronous attention-based fusion (ablation) | SODFormer · PKU-DAVIS-SOD test (all scenarios) | mAP50: 0.494 -> 0.504 | Improved |
| replace concatenation with asynchronous attention-based fusion (ablation) | SODFormer · PKU-DAVIS-SOD test (all scenarios) | mAP50: 0.478 -> 0.504 | Improved |
| replace NMS post-fusion with asynchronous attention-based fusion (ablation) | SODFormer · PKU-DAVIS-SOD test (all scenarios) | mAP50: 0.438 -> 0.504 | Improved |
| increase temporal aggregation size from 1 to 9 (ablation) | SODFormer frames · PKU-DAVIS-SOD test (all scenarios) | mAP50: 0.461 -> 0.489 | Improved |
| increase temporal aggregation size from 9 to 13 (ablation) | SODFormer frames · PKU-DAVIS-SOD test (all scenarios) | mAP50: 0.489 -> 0.488 | No clear effect |
| add DVS events to frames (ablation) | SODFormer baseline · Motion blur scenario | mAP50: 0.419 -> 0.432 | Improved |
| add DVS events to frames | SODFormer baseline · Low-light scenario | mAP50: 0.351 -> 0.374 | Improved |
| add DVS events to frames | SODFormer baseline · Normal scenario | mAP50: 0.557 -> 0.569 | Improved |
| use temporal transformer over feed-forward/recurrent detectors (frames) | SODFormer baseline · PKU-DAVIS-SOD test (all scenarios) | mAP50: 0.456 -> 0.489 | Improved |
| use temporal transformer over best event detector (events) | SODFormer baseline · PKU-DAVIS-SOD test (all scenarios) | mAP50: 0.307 -> 0.334 | Improved |
Later papers that compared with this method
1 later accepted paper reports SODFormer as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Rethinking Neuromorphic Object Detection with Hybrid Dynamic Interaction Transformers (IJCV 2026 · better on 3 of 3 shared results)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Retinomorphic Object Detection in Asynchronous Visual Streams (AAAI 2022 · similarity 0.89)
- TGBFormer: Transformer-GraphFormer Blender Network for Video Object Detection (AAAI 2025 · similarity 0.89)
- SFOD: Spiking Fusion Object Detector (CVPR 2024 · similarity 0.88)
- Scene Adaptive Sparse Transformer for Event-based Object Detection (CVPR 2024 · similarity 0.87)
- Object Detection using Event Camera: A MoE Heat Conduction based Detector and A New Benchmark Dataset (CVPR 2025 · similarity 0.86)
- Context Enhanced Transformer for Single Image Object Detection in Video Data (AAAI 2024 · 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.
