Research map / Video / Video segmentation and tracking
A Differentiable Recurrent Surface for Asynchronous Event-Based Data
Marco Cannici, Marco Ciccone, Andrea Romanoni, Matteo Matteucci
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 |
|---|---|---|---|
| replace Events-to-Video reconstruction with Matrix-LSTM | ResNet18-Ev2Vid · N-Cars | accuracy: 91.0 -> 95.80 | Improved |
| replace Events-to-Video reconstruction with Matrix-LSTM | ResNet18-Ev2Vid · N-Caltech101 | accuracy: 86.6 -> 84.12 | Hurt |
| replace hand-crafted EV-FlowNet surface with Matrix-LSTM (2 bins) | EV-FlowNet · MVSEC indoor_flying2 | AEE: 1.72 -> 1.194 | Improved |
| apply ImageNet frame normalization to Matrix-LSTM output (ablation) | ResNet18-Ev2Vid ts absolute · N-Cars | accuracy: 95.75 -> 95.22 | No clear effect |
| use ts relative instead of ts absolute time feature (ablation) | ResNet18-Ev2Vid · N-Cars | accuracy: 95.32 -> 94.77 | No clear effect |
| add global ts feature to local ts (2 bins) (ablation) | ResNet18-EST · N-Cars | accuracy: 92.75 -> 93.46 | Improved |
| add SELayer (2 bins) (ablation) | ResNet18-EST · N-Cars | accuracy: 93.46 -> 93.71 | Improved |
| increase bins from 2 to 16 (ablation) | ResNet18-EST · N-Cars | accuracy: 93.46 -> 92.01 | Hurt |
| increase output channels from 4 to 16 (1 bin) (ablation) | ResNet18-EST · N-Cars | accuracy: 93.88 -> 94.37 | Improved |
| use Matrix-LSTM instead of ConvLSTM (3x3, delay relative) (ablation) | ResNet18-Ev2Vid · N-Cars | accuracy: 92.33 -> 95.05 | Improved |
| use 2 bins instead of 1 bin (ablation) | Matrix-LSTM + EV-FlowNet · MVSEC indoor_flying2 | AEE: 1.642 -> 1.194 | Improved |
| increase bins from 2 to 4 (ablation) | Matrix-LSTM + EV-FlowNet · MVSEC indoor_flying2 | AEE: 1.194 -> 1.505 | Hurt |
Closest papers by meaning
- End-to-End Learning of Representations for Asynchronous Event-Based Data (ICCV 2019 · similarity 0.90)
- Events-To-Video: Bringing Modern Computer Vision to Event Cameras (CVPR 2019 · similarity 0.88)
- EventNet: Asynchronous Recursive Event Processing (CVPR 2019 · similarity 0.88)
- Graph Neural Network Combining Event Stream and Periodic Aggregation for Low-Latency Event-based Vision (CVPR 2025 · similarity 0.87)
- EventZoom: A Progressive Approach to Event-Based Data Augmentation for Enhanced Neuromorphic Vision (AAAI 2025 · similarity 0.86)
- Leveraging Asynchronous Spiking Neural Networks for Ultra Efficient Event-Based Visual Processing (AAAI 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.
