Research map / Label-efficient and robust learning / Recognition and long-tail
Memory Efficient Max Flow for Multi-Label Submodular MRFs
Check 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 |
|---|---|---|---|
| Store exit-flows and column-flows rather than all residual edge capacities | MEMF versus EIBFS · Middlebury stereo benchmarks | memory [MB]: 211 vs 2495; 396 vs 5907; 393 vs 5860; 219 vs 4946 | Helped |
| Store exit-flows and column-flows rather than all residual edge capacities | MEMF versus EIBFS · Cones, Teddy, KITTI, Penguin and House | memory [MB]: 1200 vs *55063; 1200 vs *55063; 2215 vs *67316; 663 vs *130728; 1986 vs *392315 | Helped |
| Use the simplified graph and flow encoding in MEMF instead of standard EIBFS | MEMF · Tsukuba, Venus, Sawtooth and Map | time [s]: 28 vs 4; 59 vs 9; 35 vs 8; 36 vs 9 | Hurt |
| Use MEMF rather than standard EIBFS | MEMF · Cones, Teddy, KITTI, Penguin and House | time [s]: 364; 2055; 18665; 6504; 9001 | Mixed |
| Replace BK with MEMF as the max-flow subroutine in IRGC+expansion | MIRGC · Penguin-128/10 and House-64/15 | memory [MB]: 4471 vs 332; 8877 vs 498 | Helped |
| Replace BK with MEMF as the max-flow subroutine in IRGC+expansion | MIRGC · Penguin-128/10 and House-64/15 | time [s]: 224 vs 2566; 106 vs 409 | Hurt |
| Use MEMF inside IRGC+expansion to handle a larger label set | MIRGC · Penguin-256/20 | memory [MB]: *17143 vs 663 | Helped |
Closest papers by meaning
- Relaxation-Based Preprocessing Techniques for Markov Random Field Inference (CVPR 2016 · similarity 0.90)
- Joint M-Best-Diverse Labelings as a Parametric Submodular Minimization (NeurIPS 2016 · similarity 0.86)
- Efficient semidefinite-programming-based inference for binary and multi-class MRFs (NeurIPS 2020 · similarity 0.83)
- Taming the Sigmoid Bottleneck: Provably Argmaxable Sparse Multi-Label Classification (AAAI 2024 · similarity 0.83)
- Markovian Flow Matching: Accelerating MCMC with Continuous Normalizing Flows (NeurIPS 2024 · similarity 0.83)
- Learning Juntas under Markov Random Fields (NeurIPS 2025 · similarity 0.82)
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.