Research map / 3D vision / Depth and stereo
WiFlow: Estimating Optical Flow using WiFi Channel State Information
Thomas Weigel, Simon Kiefhaber, Fabian Portner, Matthias Hollick, Simone Schaub-Meyer
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 |
|---|---|---|---|
| use Quotient (antenna ratio) preprocessing instead of Raw CSI (ablation) | WiFlow flow-only · WiFlow birdview+ time split | EPE_M: 3.71 -> 2.89 | Improved |
| use Quotient preprocessing instead of Raw CSI (ablation) | WiFlow flow-only · WiFlow birdview+ time split | EPE_A: 0.94 -> 0.78 | Improved |
| use PCA denoising instead of Raw CSI (ablation) | WiFlow flow-only · WiFlow birdview+ time split | EPE_M: 3.71 -> 2.95 | Improved |
| use Savitzky-Golay smoothing instead of Raw CSI (ablation) | WiFlow flow-only · WiFlow birdview+ time split | EPE_A: 0.94 -> 1.14 | Hurt |
| use Fourier transform instead of Raw CSI (ablation) | WiFlow flow-only · WiFlow birdview+ time split | EPE_M: 3.71 -> 3.71 | No clear effect |
| train a model instead of predicting zero flow (Raw preprocessing, plain EPE) (ablation) | WiFlow flow-only · WiFlow birdview+ time split | EPE: 0.17 -> 0.40 | Hurt |
| add more receiver devices (1 to 4) (ablation) | WiFlow flow-only · WiFlow birdview+ | EPE metrics | Improved |
| add parallel mask block (flow x mask) to flow-only model | WiFlow · WiFlow birdview+ all splits | EPE metrics | Mixed |
| add mask-based localization to flow-only model | WiFlow · off-area and void actions | EPE | Improved |
| double target flow resolution (128x168 to 256x336) (ablation) | WiFlow models · WiFlow birdview+ subject split | EPE | Hurt |
| evaluate on unseen subjects (subject split vs time split) | WiFlow models · WiFlow birdview+ | EPE | No clear effect |
| use parallel flow+mask architecture instead of flow-only (compute cost) | WiFlow · WiFlow birdview+ inference | inference time (ms): 23 -> 47 | Hurt |
Closest papers by meaning
- VideoFlow: Exploiting Temporal Cues for Multi-frame Optical Flow Estimation (ICCV 2023 · similarity 0.86)
- StreamFlow: Streamlined Multi-Frame Optical Flow Estimation for Video Sequences (NeurIPS 2024 · similarity 0.86)
- MemFlow: Optical Flow Estimation and Prediction with Memory (CVPR 2024 · similarity 0.84)
- Single Image Optical Flow Estimation With an Event Camera (CVPR 2020 · similarity 0.84)
- DPFlow: Adaptive Optical Flow Estimation with a Dual-Pyramid Framework (CVPR 2025 · similarity 0.84)
- FlowTrack: Revisiting Optical Flow for Long-Range Dense Tracking (CVPR 2024 · similarity 0.83)
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.
