Research map / Video / Video segmentation and tracking
Continuous Parametric Optical Flow
Read the paperCheck 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 |
|---|---|---|---|
| add neural ODE unit to feature encoder (vs basic RAFT encoder) (ablation) | ODE-6spline · Kubric Query-Stride 8f | ADE_All: 4.89 -> 3.77 | Helped |
| add neural ODE unit to feature encoder (ablation) | ODE-6spline · Kubric Query-Stride 24f | ADE_All: 9.92 -> 8.86 | Helped |
| add neural ODE unit to feature encoder (ablation) | ODE-6spline · Kubric Query-Stride 8f | TRMSE: 5.45 -> 4.26 | Helped |
| use 6 instead of 4 B-spline control points (ablation) | ODE-6spline · Kubric Query-Stride 8f | ADE_All: 4.41 -> 3.77 | Helped |
| use 6 instead of 4 B-spline control points (ablation) | ODE-6spline · Kubric Query-Stride 24f | TRMSE: 12.65 -> 12.03 | Helped |
| use 6 instead of 8 B-spline control points (ablation) | ODE-6spline · Kubric Query-Stride 8f | ADE_All: 3.94 -> 3.77 | Helped |
| use 6 instead of 8 B-spline control points (ablation) | ODE-6spline · Kubric Query-Stride 24f | ADE_All: 9.13 -> 8.86 | Helped |
| replace PIPs with continuous parametric model | ODE-6spline · Kubric Query-Stride 24f | ADE_All: 12.09 -> 10.61 | Helped |
| replace RAFT with continuous parametric model | ODE-6spline · Kubric Query-Stride 24f | ADE_All: 16.37 -> 10.61 | Helped |
| replace PIPs with continuous parametric model | ODE-6spline · TAP-Vid-Kinetics 250f | ADE_All: 40.59 -> 34.42 | Helped |
| replace PIPs with continuous parametric model | ODE-6spline · TAP-Vid-DAVIS 20f | ADE_All: 15.18 -> 13.48 | Helped |
| replace PIPs with continuous parametric model | ODE-6spline · Kubric Query-Stride 10f | ADE_Occ (sampling): 7.71 -> 8.09 | Hurt |
Closest papers by meaning
- Recurrent Partial Kernel Network for Efficient Optical Flow Estimation (AAAI 2024 · similarity 0.85)
- FlowFeat: Pixel-Dense Embedding of Motion Profiles (NeurIPS 2025 · similarity 0.84)
- TimeTracker: Event-based Continuous Point Tracking for Video Frame Interpolation with Non-linear Motion (CVPR 2025 · similarity 0.84)
- EDCFlow: Exploring Temporally Dense Difference Maps for Event-based Optical Flow Estimation (CVPR 2025 · similarity 0.84)
- FlowTrack: Revisiting Optical Flow for Long-Range Dense Tracking (CVPR 2024 · similarity 0.84)
- Dense Optical Tracking: Connecting the Dots (CVPR 2024 · similarity 0.84)
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.