Research map / Video / Video segmentation and tracking
Boosting Text-Driven Video Segmentation via Geometry-Aware Distillation
Tianyu Zhu, Yingping Liang, Hesong Li, Ying Fu
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 monocular geometry pretraining (MGP, Stage I) (ablation) | GeoLaV (SAMWISE base) · Ref-YouTube-VOS fine-tuned | J&F: 65.1 -> 67.2 | Improved |
| add geometry-aware distillation (GAD, Stage II) (ablation) | GeoLaV (SAMWISE base) · Ref-YouTube-VOS fine-tuned | J&F: 65.1 -> 67.5 | Improved |
| combine MGP and GAD (ablation) | GeoLaV · Ref-YouTube-VOS fine-tuned | J&F: 67.5 -> 70.5 | Improved |
| add planar homography augmentation instead of geometry (ablation) | GeoLaV (SAMWISE base) · Ref-YouTube-VOS fine-tuned | J&F: 65.1 -> 66.5 | Improved |
| replace planar augmentation with depth-based MGP (ablation) | GeoLaV · Ref-YouTube-VOS fine-tuned | J&F: 66.5 -> 67.2 | Improved |
| add MGP to image-only pretraining, zero-shot to video | GeoLaV · Ref-YouTube-VOS zero-shot | J&F: 31.9 -> 47.0; +15.1 | Improved |
| add MGP to image-only pretraining, zero-shot to video | GeoLaV · MeViS zero-shot | J&F: 26.4 -> 31.6; +5.2 | Improved |
| full GeoLaV versus SAMWISE (202M) | GeoLaV · Ref-YouTube-VOS | J&F: 69.2 -> 70.5 | Improved |
| full GeoLaV versus SAMWISE (202M) | GeoLaV · Ref-DAVIS17 | J&F: 70.6 -> 72.5 | Improved |
| full GeoLaV versus SAMWISE (202M) | GeoLaV · MeViS | J&F: 49.5 -> 50.0 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- SAMWISE (202M) — SAMWISE: Infusing Wisdom in SAM2 for Text-Driven Video Segmentation (CVPR 2025) · better on 6 of 6 shared results
- ReferDINO — ReferDINO: Referring Video Object Segmentation with Visual Grounding Foundations (ICCV 2025) · better on 3 of 3 shared results
- VISA-7B — VISA: Reasoning Video Object Segmentation via Large Language Model (ECCV 2024) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Deforming Videos to Masks: Flow Matching for Referring Video Segmentation (ICLR 2026 · similarity 0.90)
- End-to-End Referring Video Object Segmentation With Multimodal Transformers (CVPR 2022 · similarity 0.88)
- Semantic and Sequential Alignment for Referring Video Object Segmentation (CVPR 2025 · similarity 0.88)
- LoSh: Long-Short Text Joint Prediction Network for Referring Video Object Segmentation (CVPR 2024 · similarity 0.88)
- SAMWISE: Infusing Wisdom in SAM2 for Text-Driven Video Segmentation (CVPR 2025 · similarity 0.87)
- Learning Spatial-Semantic Features for Robust Video Object Segmentation (ICLR 2025 · similarity 0.87)
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.
