Research map / 3D vision / Depth and stereo
Changes in Real Time: Online Scene Change Detection with Multi-View Fusion
Chamuditha Jayanga Galappaththige, Jason Lai, Lloyd Windrim, Donald Dansereau, Niko Suenderhauf, Dimity Miller
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 L1 term to pixel change cue (ablation) | Online SCD · PASLCD | mIoU: 0.320 -> 0.486 | Improved |
| add D-SSIM term to pixel change cue (ablation) | Online SCD · PASLCD | mIoU: 0.447 -> 0.486 | Improved |
| use self-supervised fusion loss instead of hard thresholding and intersection (ablation) | Online SCD · PASLCD | mIoU: 0.350 -> 0.486 | Improved |
| use both pixel and feature change cues instead of one (ablation) | Online SCD · PASLCD | mIoU | Improved |
| add regularization term to fusion loss (ablation) | Online SCD · PASLCD | mIoU | Improved |
| restrict global optimization to changed primitives (ablation) | Scene update · PASLCD | Runtime (s): 145 -> 79 | Improved |
| add selective reconstruction to global optimization (ablation) | Scene update · PASLCD | PSNR (dB): 23.01 -> 23.70 | Improved |
| add selective reconstruction to global optimization (ablation) | Scene update · PASLCD | Runtime (s): 79 -> 36 | Improved |
| add global optimization to selective reconstruction (ablation) | Scene update · PASLCD | PSNR (dB): 19.89 -> 23.70 | Improved |
| reduce multi-view fusion iterations (ablation) | Online SCD · PASLCD | F1 / FPS: 11-20 FPS, 3.6% F1 drop | Mixed |
| use change-guided update instead of retraining 3DGS from scratch | Scene update · PASLCD | PSNR (dB): 22.21 -> 23.70 | Improved |
| online multi-view fusion method vs best offline MV3DCD | Ours online · PASLCD | mIoU: 0.478 -> 0.486 | Improved |
Closest papers by meaning
- Multi-View Pose-Agnostic Change Localization with Zero Labels (CVPR 2025 · similarity 0.89)
- Towards Generalizable Scene Change Detection (CVPR 2025 · similarity 0.89)
- EMPLACE: Self-Supervised Urban Scene Change Detection (AAAI 2025 · similarity 0.86)
- Zero-Shot Scene Change Detection (AAAI 2025 · similarity 0.86)
- Living Scenes: Multi-object Relocalization and Reconstruction in Changing 3D Environments (CVPR 2024 · similarity 0.85)
- The STVchrono Dataset: Towards Continuous Change Recognition in Time (CVPR 2024 · similarity 0.85)
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.
