Research map / Detection and segmentation / Open-set and few-shot detection
ESS: Learning Event-Based Semantic Segmentation from Still Images
Zhaoning Sun, Nico Messikommer, Daniel Gehrig, Davide Scaramuzza
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 consistency loss in embedding space (ablation) | ESS · DSEC-Semantic (UDA, no event labels) | mIoU (%): 39.31 -> 44.87 | Improved |
| add consistency loss on predictions (ablation) | ESS · DSEC-Semantic (UDA, no event labels) | mIoU (%): 43.59 -> 44.87 | Improved |
| add consistency loss on task outputs (ablation) | ESS · DSEC-Semantic (UDA, no event labels) | mIoU (%): 42.78 -> 44.87 | Improved |
| add skip connections (ablation) | ESS · DSEC-Semantic (UDA, no event labels) | mIoU (%): 38.08 -> 44.87 | Improved |
| add embedding-space consistency loss (ablation) | ESS · DSEC-Semantic (UDA, no event labels) | accuracy (%): 80.86 -> 84.04 | Improved |
| ESS instead of VID2E in UDA setting | ESS · DDD17 (UDA, no event labels) | mIoU (%): 45.48 -> 52.46 | Improved |
| ESS instead of E2VID in UDA setting | ESS · DSEC-Semantic (UDA, no event labels) | mIoU (%): 40.70 -> 44.87 | Improved |
| add image labels to event labels in training | ESS · DSEC-Semantic (supervised) | mIoU (%): 51.57 -> 53.29 | Improved |
| add image labels to event labels in training | ESS · DDD17 (supervised, event labels) | mIoU (%): 61.37 -> 60.43 | Hurt |
| ESS with only event labels versus EV-SegNet | ESS · DSEC-Semantic (supervised) | mIoU (%): 51.76 -> 51.57 | No clear effect |
| ESS instead of E2VID for embedding alignment | ESS · DSEC-Semantic embedding alignment (UDA) | symmetric KL dissimilarity: 0.073 -> 0.025 | Improved |
Closest papers by meaning
- ESEG: Event-Based Segmentation Boosted by Explicit Edge-Semantic Guidance (AAAI 2025 · similarity 0.93)
- OpenESS: Event-based Semantic Scene Understanding with Open Vocabularies (CVPR 2024 · similarity 0.92)
- Efficient Event-Based Semantic Segmentation via Exploiting Frame-Event Fusion: A Hybrid Neural Network Approach (AAAI 2025 · similarity 0.92)
- HPL-ESS: Hybrid Pseudo-Labeling for Unsupervised Event-based Semantic Segmentation (CVPR 2024 · similarity 0.90)
- LaSe-E2V: Towards Language-guided Semantic-aware Event-to-Video Reconstruction (NeurIPS 2024 · similarity 0.89)
- Event-assisted Low-Light Video Object Segmentation (CVPR 2024 · similarity 0.89)
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.
