Research map / Video / Video segmentation and tracking
Learning To Recommend Frame for Interactive Video Object Segmentation in the Wild
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 learned frame recommendation instead of worst-quality frame selection | Interactive VOS with IPN · DAVIS (Oracle) | AUC: 48.02 -> 48.25 | Helped |
| Use learned frame recommendation instead of worst-quality frame selection | Interactive VOS with IPN · YouTube-VOS (Oracle) | AUC: 44.67 -> 43.86 | Hurt |
| Use learned frame recommendation instead of the strongest listed Wild baseline | Interactive VOS with ATNet · DAVIS (Wild) | AUC: 73.3266 -> 74.1047 | Helped |
| Use learned frame recommendation instead of the strongest listed Wild baseline | Interactive VOS with ATNet · YouTube-VOS (Wild) | AUC: 74.6877 -> 75.33 | Helped |
| Use a goal-reward threshold of mean plus standard deviation rather than mean (ablation) | Frame recommendation agent · DAVIS validation (Oracle) | AUC: 71.8211 -> 74.01 | Helped |
| Add the auxiliary diversity reward (ablation) | Frame recommendation agent · DAVIS validation (Oracle) | AUC: 73.7465 -> 74.01 | Helped |
| Add the goal reward (ablation) | Frame recommendation agent · DAVIS validation (Oracle) | AUC: 72.0605 -> 74.01 | Helped |
| Include recommendation history in the state alongside segmentation quality (ablation) | Frame recommendation agent · DAVIS validation (Oracle) | AUC: 73.7088 -> 74.01 | Helped |
| Include segmentation quality in the state alongside recommendation history (ablation) | Frame recommendation agent · DAVIS validation (Oracle) | AUC: 73.9151 -> 74.01 | Helped |
| Use task decomposition during training (ablation) | Frame recommendation agent · DAVIS validation (Oracle) | AUC: 72.5831 -> 74.01 | Helped |
| Use learned frame recommendation rather than human frame selection | Interactive VOS with ATNet · DAVIS | AUC: 73.09 -> 74.1047 | Helped |
| Use learned frame recommendation rather than human frame selection | Interactive VOS with ATNet · DAVIS | time (seconds): 14.01 -> 0.70 | Helped |
Closest papers by meaning
- Reinforcement Cutting-Agent Learning for Video Object Segmentation (CVPR 2018 · similarity 0.88)
- Fast User-Guided Video Object Segmentation by Interaction-And-Propagation Networks (CVPR 2019 · similarity 0.87)
- Modular Interactive Video Object Segmentation: Interaction-to-Mask, Propagation and Difference-Aware Fusion (CVPR 2021 · similarity 0.87)
- BubbleNets: Learning to Select the Guidance Frame in Video Object Segmentation by Deep Sorting Frames (CVPR 2019 · similarity 0.87)
- Siamese Network with Interactive Transformer for Video Object Segmentation (AAAI 2022 · similarity 0.87)
- Point-VOS: Pointing Up Video Object Segmentation (CVPR 2024 · 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.