Research map / Graph learning / Graph self-supervised learning
Visibility Graphs for Image Processing
Jacopo Iacovacci, Lucas Lacasa
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 local visibility patch features to global degree distribution (ablation) | IVG+IHVG SVM · 2D HeLa | accuracy: 75.4% -> 83.5% | Improved |
| add global degree distribution to local patch profiles (ablation) | IVG+IHVG SVM · 2D HeLa | accuracy: 79.5% -> 83.5% | Improved |
| add global and local multiplex features together (vs global only) (ablation) | multiplex IVG+IHVG ova-qSVM · Usptex | accuracy: 93.6% -> 98.4% | Improved |
| add global and local multiplex features together (vs local only) (ablation) | multiplex IVG+IHVG ova-qSVM · Multiband | accuracy: 99.4% -> 99.6% | Improved |
| add global and local multiplex features together (vs global only) (ablation) | multiplex IVG+IHVG ova-qSVM · Brodatz Colored | accuracy: 98.7% -> 99.8% | Improved |
| use IHVG instead of IVG for global degree distribution | P(k) features · 2D HeLa | model average accuracy: 57.2% -> 67.6% | Improved |
| use IVG instead of IHVG for local patch profile | Z features · 2D HeLa | model average accuracy: 74% -> 75.8% | Improved |
| add PCA to IVG P(k) features (ablation) | quadratic SVM / weighted KNN · Kylberg | accuracy: 94.2% -> 100% | Mixed |
| use IHVG instead of IVG degree distribution | Complex Tree · Kylberg | accuracy: 95.3% -> 100% | Improved |
| apply k-filter degree cut-off at 24 | IHVG k-filter compressor · Lena 512x512 | file size: 50% size reduction | Improved |
| use C-filter instead of k-filter at high compression | IHVG filters · Lena | IC ratio IC_F(I)/IC_I: 0.2 at FCP=1.5 vs 0.5 at FCP=2 for k-filter | Hurt |
| use C-filter instead of k-filter with no compression | IHVG filters · Lena | IC ratio IC_F(I)/IC_I: 0.97 at FCP=1 | Improved |
Closest papers by meaning
- VisDiff: SDF-Guided Polygon Generation for Visibility Reconstruction, Characterization and Recognition (NeurIPS 2025 · similarity 0.84)
- Vision HGNN: An Image is More than a Graph of Nodes (ICCV 2023 · similarity 0.83)
- Fast Tree-Field Integrators: From Low Displacement Rank to Topological Transformers (NeurIPS 2024 · similarity 0.82)
- Hypergraph Vision Transformers: Images are More than Nodes, More than Edges (CVPR 2025 · similarity 0.82)
- Image Processing GNN: Breaking Rigidity in Super-Resolution (CVPR 2024 · similarity 0.82)
- GreedyViG: Dynamic Axial Graph Construction for Efficient Vision GNNs (CVPR 2024 · similarity 0.81)
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.
