Research map / Label-efficient and robust learning / Recognition and long-tail
MRF Optimization with Separable Convex Prior on Partially Ordered Labels
Csaba Domokos, Frank R. Schmidt, Daniel Cremers
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 proposed coarse-to-fine separable convex prior method instead of FullFlow at 64x64 labels | optical flow MRF optimization · MPI Sintel ambush 7 (64x64) | runtime (s): 47.33 -> 0.89 | Improved |
| use proposed method instead of FullFlow at 64x64 labels | optical flow MRF optimization · MPI Sintel ambush 7 (64x64) | EPE: 0.65 -> 1.40 | Hurt |
| use proposed method instead of FullFlow at 8x8 labels | optical flow MRF optimization · MPI Sintel sleeping 1 (8x8) | runtime (s): 0.86 -> 0.35 | Improved |
| use proposed method instead of alpha-expansion at 64x64 labels | optical flow MRF optimization · MPI Sintel ambush 7 (64x64) | runtime (s): 286.73 -> 0.89 | Improved |
| use proposed method instead of alpha-expansion at 64x64 labels | optical flow MRF optimization · MPI Sintel ambush 7 (64x64) | EPE: 1.36 -> 1.40 | Hurt |
| use proposed method instead of Shekhovtsov et al. [12] | optical flow MRF optimization · MPI Sintel ambush 7 (64x64) | EPE: 3.30 -> 1.40 | Improved |
| use proposed method instead of alternating baseline at 8x8 labels | optical flow MRF optimization · MPI Sintel sleeping 1 (8x8) | EPE: 1.12 -> 0.60 | Improved |
| grow label set from 8x8 to 64x64 for proposed method | Proposed · MPI Sintel | EPE: 0.60 -> 1.40 | Hurt |
| add label refinement with local alpha-beta swap on 3x3 label neighborhoods (ablation) | Proposed · MPI Sintel | EPE and runtime | Mixed |
Closest papers by meaning
- Sublabel-Accurate Convex Relaxation of Vectorial Multilabel Energies (ECCV 2016 · similarity 0.86)
- Sublabel-Accurate Relaxation of Nonconvex Energies (CVPR 2016 · similarity 0.86)
- A Cutting-Plane Method for Sublabel-Accurate Relaxation of Problems with Product Label Spaces (IJCV 2022 · similarity 0.86)
- An Inference Algorithm for Multi-Label MRF-MAP Problems with Clique Size 100 (ECCV 2020 · similarity 0.86)
- A Discriminative View of MRF Pre-Processing Algorithms (ICCV 2017 · similarity 0.85)
- New Convex Relaxations for MRF Inference With Unknown Graphs (ICCV 2019 · 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.
