Research map / 3D vision / Depth and stereo
MV-TAP: Tracking Any Point in Multi-View Videos
Jahyeok Koo, Inès Hyeonsu Kim, Mungyeom Kim, Junghyun Park, Seohyeon Park, Jaeyeong Kim, Jung Yi, Seokju Cho, Seungryong Kim
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 view attention and camera embedding to CoTracker3 (MV-TAP) (ablation) | CoTracker3 · DexYCB (8 views) | AJ: 41.5 -> 44.2 | Improved |
| add view attention to CoTracker3 (ablation) | CoTracker3 · Panoptic Studio | AJ: 39.6 -> 38.6 | Hurt |
| add view attention to CoTracker3 (ablation) | CoTracker3 · DexYCB | AJ: 41.5 -> 43.6 | Improved |
| add camera embedding to view attention (ablation) | MV-TAP · Panoptic Studio | AJ: 38.6 -> 40.3 | Improved |
| add camera embedding to view attention (ablation) | MV-TAP · DexYCB | AJ: 43.6 -> 44.2 | Improved |
| increase number of input views from 2 to 8 (ablation) | MV-TAP · DexYCB | AJ: 39.2 -> 44.2 | Improved |
| increase number of input views from 2 to 8 (ablation) | CoTracker3 · DexYCB | AJ: 37.5 -> 41.5 | Improved |
| flatten view and time dimensions in CoTracker3 | CoTracker3 · DexYCB (8 views) | AJ: 41.5 -> 2.7 | Hurt |
| add triangulation-based refinement to CoTracker3 | CoTracker3 · DexYCB (in-frame occluded points) | position accuracy on occluded points: 33.9 -> 34.8 | Improved |
| add triangulation-based refinement to CoTracker3 | CoTracker3 · DexYCB (8 views) | position accuracy (<d_avg): 59.6 -> 57.1 | Hurt |
| MV-TAP versus additionally trained CoTracker3 (ablation) | MV-TAP · DexYCB | AJ: 41.8 -> 44.2 | Improved |
| MV-TAP on frequently occluded trajectories | CoTracker3 · DexYCB (frequently occluded trajectories) | AJ: 26.2 -> 29.7 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- TAPIP3D — TAPIP3D: Tracking Any Point in Persistent 3D Geometry (NeurIPS 2025) · better on 2 of 4 shared results
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
- TAPVid-3D: A Benchmark for Tracking Any Point in 3D (NeurIPS 2024 · similarity 0.89)
- TAPIP3D: Tracking Any Point in Persistent 3D Geometry (NeurIPS 2025 · similarity 0.88)
- ETAP: Event-based Tracking of Any Point (CVPR 2025 · similarity 0.87)
- TAPVid-360: Tracking Any Point in 360 from Narrow Field of View Video (NeurIPS 2025 · similarity 0.87)
- Tracking Everything Everywhere across Multiple Cameras (AAAI 2025 · similarity 0.87)
- Tracktention: Leveraging Point Tracking to Attend Videos Faster and Better (CVPR 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.
