Research map / Detection and segmentation / Open-set and few-shot detection
Memory Aggregation Networks for Efficient Interactive Video Object Segmentation
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 |
|---|---|---|---|
| Use MA-Net instead of the strongest reported AUC baseline, Heo et al. with optical flow | Interactive video object segmentation · DAVIS 2017 validation | AUC: 0.704 -> 0.749 | Helped |
| Use MA-Net instead of the strongest J@60 baseline, Oh et al. | Interactive video object segmentation · DAVIS 2017 validation | J@60: 0.734 -> 0.761 | Helped |
| Add global memory to aggregate matching maps across interaction rounds (ablation) | MA-Net · DAVIS 2017 validation | Jaccard score | Helped |
| Add local memory to reuse local matching maps from earlier rounds (ablation) | MA-Net · DAVIS 2017 validation | Jaccard score | Helped |
| Add both global and local memory (ablation) | MA-Net · DAVIS 2017 validation | Jaccard score | Helped |
| Limit local-map memory to the current round (R=1) (ablation) | MA-Net · DAVIS 2017 validation | Segmentation accuracy | Hurt |
| Use local maps from the past two rounds (R=2) (ablation) | MA-Net · DAVIS 2017 validation | Segmentation accuracy | Helped |
| Remove the augmented map from global memory (ablation) | MA-Net · DAVIS 2017 validation | AUC: 0.749 -> 0.744 | Hurt |
| Set local window size k to 6 instead of 12 (ablation) | MA-Net · DAVIS 2017 validation | AUC: 0.749 -> 0.724 | Hurt |
| Set local window size k to 9 instead of 12 (ablation) | MA-Net · DAVIS 2017 validation | AUC: 0.749 -> 0.737 | Hurt |
| Set local window size k to 15 instead of 12 (ablation) | MA-Net · DAVIS 2017 validation | AUC: 0.749 -> 0.748 | No clear effect |
Closest papers by meaning
- Query-Memory Re-Aggregation for Weakly-supervised Video Object Segmentation (AAAI 2021 · similarity 0.88)
- Fast User-Guided Video Object Segmentation by Interaction-And-Propagation Networks (CVPR 2019 · similarity 0.88)
- Recurrent Dynamic Embedding for Video Object Segmentation (CVPR 2022 · similarity 0.88)
- Putting the Object Back into Video Object Segmentation (CVPR 2024 · similarity 0.87)
- RMem: Restricted Memory Banks Improve Video Object Segmentation (CVPR 2024 · similarity 0.86)
- LiVOS: Light Video Object Segmentation with Gated Linear Matching (CVPR 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.