Research map / Video / Video segmentation and tracking
EVOLVE: Event-Guided Deformable Feature Transfer and Dual-Memory Refinement for Low-Light Video Object Segmentation
Jong-Hyeon Baek, Jiwon Oh, Yeong Jun Koh
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 event-guided deformable feature transfer (EDFT) and dual-memory object transformer (DMOT) (ablation) | EVOLVE · LLE-VOS ablation | J&F: 0.578 -> 0.773 | Improved |
| predict deformable offsets from event features instead of image features (ablation) | EDFT · LLE-VOS ablation | J&F: 0.724 -> 0.773 | Improved |
| apply deformable convolution to image features rather than adding it to event features too (ablation) | EDFT · LLE-VOS ablation | J&F: 0.757 -> 0.773 | Improved |
| use dual image and event memories instead of single memory (ablation) | DMOT · LLE-VOS ablation | J&F: 0.740 -> 0.773 | Improved |
| add memory refinement module (MRM) that updates memories iteratively (ablation) | DMOT · LLE-VOS ablation | J&F: 0.742 -> 0.773 | Improved |
| add IoU-based confidence weighting of memory tokens (ablation) | MRM · LLE-VOS ablation | J&F: 0.754 -> 0.773 | Improved |
| use MRM instead of mask pooling from Cutie (ablation) | DMOT · LLE-VOS ablation | J&F: 0.762 -> 0.773 | Improved |
| fuse image and event with EVOLVE instead of simple concatenation (ablation) | EVOLVE · LLE-VOS ablation | J&F: 0.710 -> 0.773 | Improved |
| use event-only input without images (ablation) | EVOLVE · LLE-VOS ablation | J&F: 0.773 -> 0.627 | Hurt |
Closest papers by meaning
- Event-assisted Low-Light Video Object Segmentation (CVPR 2024 · similarity 0.91)
- Breaking the "Object" in Video Object Segmentation (CVPR 2023 · similarity 0.87)
- Domain Generalization for Multiple Video Object Segmentation and Tracking Using Transformers and Smart Memory (IJCV 2026 · similarity 0.86)
- Video State-Changing Object Segmentation (ICCV 2023 · similarity 0.86)
- LiVOS: Light Video Object Segmentation with Gated Linear Matching (CVPR 2025 · similarity 0.85)
- Referred by Multi-Modality: A Unified Temporal Transformer for Video Object Segmentation (AAAI 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.
