Research map / Video / Video segmentation and tracking
Video to Events: Recycling Video Datasets for Event Cameras
Check 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 |
|---|---|---|---|
| Add contrast-threshold randomization during synthetic event generation (ablation) | Synthetic-event ResNet-34 classifier · N-Caltech101 | accuracy: 0.751 -> 0.782 | Helped |
| Add dataset extension with additional images for classes (ablation) | Synthetic-event ResNet-34 classifier · N-Caltech101 | accuracy: 0.751 -> 0.769 | Helped |
| Add both contrast-threshold randomization and dataset extension (ablation) | Synthetic-event ResNet-34 classifier · N-Caltech101 | accuracy: 0.751 -> 0.807 | Helped |
| Fine-tune synthetic-event training on real events (ablation) | Synthetic-event ResNet-34 classifier · N-Caltech101 | accuracy: 0.807 -> 0.906 | Helped |
| Add contrast-threshold randomization to real-event fine-tuning (ablation) | Synthetic-event ResNet-34 classifier · N-Caltech101 | accuracy: 0.856 -> 0.904 | Helped |
| Add dataset extension to real-event fine-tuning (ablation) | Synthetic-event ResNet-34 classifier · N-Caltech101 | accuracy: 0.856 -> 0.852 | No clear effect |
| Apply frame interpolation to video downsampled by factor 80 (ablation) | Synthetic event generation and ResNet-34 classification · N-Caltech101, low-rate video factor 80 | test score: 0.618 -> 0.687 | Helped |
| Use synthetic events instead of real events for training | Semantic segmentation with EST, 50 ms · DDD17 test split | MIoU: 54.81 -> 45.48 | Hurt |
| Fine-tune synthetic-event EST segmentation model on real events (ablation) | Semantic segmentation with EST, 50 ms · DDD17 test split | MIoU: 45.48 -> 56.01 | Helped |
| Use EST rather than the Alonso et al. event representation after fine-tuning (ablation) | U-Net with Xception encoder · DDD17 test split, 50 ms | MIoU: 55.17 -> 56.01 | Helped |
| Fine-tune Alonso-representation segmentation model on real events (ablation) | Synthetic-event segmentation model · DDD17 test split, 50 ms | MIoU: 48.16 -> 55.17 | Helped |
| Train on synthetic events rather than real events | Alonso-representation segmentation model · DDD17 test split, 50 ms | accuracy: 89.76 -> 86.03 | Hurt |
Closest papers by meaning
- E2HQV: High-Quality Video Generation from Event Camera via Theory-Inspired Model-Aided Deep Learning (AAAI 2024 · similarity 0.89)
- EventZoom: A Progressive Approach to Event-Based Data Augmentation for Enhanced Neuromorphic Vision (AAAI 2025 · similarity 0.88)
- Event-assisted Low-Light Video Object Segmentation (CVPR 2024 · similarity 0.87)
- Efficient Event-Based Semantic Segmentation via Exploiting Frame-Event Fusion: A Hybrid Neural Network Approach (AAAI 2025 · similarity 0.87)
- V2V: Scaling Event-Based Vision through Efficient Video-to-Voxel Simulation (NeurIPS 2025 · similarity 0.87)
- LaSe-E2V: Towards Language-guided Semantic-aware Event-to-Video Reconstruction (NeurIPS 2024 · similarity 0.87)
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.