Research map / Video / Video segmentation and tracking
Simultaneous Motion And Noise Estimation with Event Cameras
Shintaro Shiba, Yoshimitsu Aoki, Guillermo Gallego
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 joint noise estimation (keep 90% events) to CMax | CMax · ECD boxes_rot, 5 Hz noise | RMS: 124.641377 -> 97.58457484 | Improved |
| add joint noise estimation to CMax | CMax · ECD dynamic_rot, 5 Hz noise | RMS: 20.00070968 -> 8.521574073 | Improved |
| replace random downsampling with proposed denoising | CMax · ECD boxes_rot, 5 Hz noise | RMS: 124.8686397 -> 97.58457484 | Improved |
| add good initialization from previous slice | CMax · ECD dynamic_rot, 5 Hz noise | RMS: 20.00070968 -> 8.274876245 | Improved |
| lower kept-signal ratio tau from 90% to 70% (ablation) | Ours · ECD boxes_rot, 5 Hz noise | RMS: 97.58457484 -> 163.6762405 | Hurt |
| add joint denoising to MultiCM flow estimation | MultiCM · MVSEC indoor1, 5 Hz noise | EPE: 2.675668117 -> 2.516871532 | Improved |
| add joint denoising to MultiCM flow estimation | MultiCM · MVSEC outdoor1, 5 Hz noise | EPE: 1.960064541 -> 1.86462922 | Improved |
| swap MultiCM for deep E-RAFT as motion estimator (ablation) | Ours · DSEC zurich_city_12a | ESR: 2.45 -> 2.39 | No clear effect |
| use unsupervised joint estimation against supervised EDformer | Ours · DND21 hotel 1Hz | AUC: 0.9928 vs 1.0138021943171762 | Hurt |
| add denoising overhead | Ours · 30k events ECD runtime | time: 0.1 s of 2 s | No clear effect |
Closest papers by meaning
- Secrets of Event-Based Optical Flow, Depth and Ego-Motion Estimation by Contrast Maximization (TPAMI 2024 · similarity 0.91)
- LED: A Large-scale Real-world Paired Dataset for Event Camera Denoising (CVPR 2024 · similarity 0.90)
- Secrets of Event-Based Optical Flow (ECCV 2022 · similarity 0.90)
- ProgressiveMotionSeg: Mutually Reinforced Framework for Event-Based Motion Segmentation (AAAI 2022 · similarity 0.89)
- Event-Based Motion Segmentation by Motion Compensation (ICCV 2019 · similarity 0.89)
- A Unifying Contrast Maximization Framework for Event Cameras, With Applications to Motion, Depth, and Optical Flow Estimation (CVPR 2018 · similarity 0.88)
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.
