Research map / Learning theory and algorithms / Clustering
FLIC: Fast Linear Iterative Clustering With Active Search
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 |
|---|---|---|---|
| Use active search with local neighboring-pixel continuity | FLIC · BSDS500 | boundary recall | Helped |
| Use active search with local neighboring-pixel continuity | FLIC · BSDS500 | time: 0.035s vs 0.625s | Helped |
| Use 8-neighborhood instead of 4-neighborhood (ablation) | FLIC · BSDS500, K=200 | boundary recall (%): 85.9 -> 87.4 | Helped |
| Use 8-neighborhood instead of 4-neighborhood (ablation) | FLIC · BSDS500, K=200 | time: 35 -> 56 | Hurt |
| Use back-and-forth scans instead of repeated forward scans (ablation) | FLIC · BSDS500 | boundary recall | Helped |
| Use one iteration instead of two (ablation) | FLIC · BSDS500, K=200 | boundary recall: 0.835 -> 0.859 | Hurt |
| Use three iterations instead of two (ablation) | FLIC · BSDS500, K=200 | boundary recall: 0.859 -> 0.860 | Helped |
| Use two iterations instead of one (ablation) | FLIC · BSDS500, K=200 | undersegment error: 0.115 -> 0.108 | Helped |
| Use three iterations instead of two (ablation) | FLIC · BSDS500, K=200 | undersegment error: 0.108 -> 0.107 | Helped |
| Use two iterations instead of one (ablation) | FLIC · BSDS500, K=200 | achievable segmentation accuracy: 0.941 -> 0.945 | Helped |
| Use three iterations instead of two (ablation) | FLIC · BSDS500, K=200 | achievable segmentation accuracy: 0.945 -> 0.946 | Helped |
| Perform assignment and seed update jointly rather than separately (ablation) | FLIC · BSDS500 | iterations to reach the same boundary recall: 2 fewer iterations | Helped |
Closest papers by meaning
- Superpixels and Polygons Using Simple Non-Iterative Clustering (CVPR 2017 · similarity 0.88)
- CLUSTSEG: Clustering for Universal Segmentation (ICML 2023 · similarity 0.85)
- Image Clustering Conditioned on Text Criteria (ICLR 2024 · similarity 0.83)
- Hierarchical Overlapping Clustering on Graphs: Cost Function, Algorithm and Scalability (ICML 2025 · similarity 0.81)
- Learning Hierarchical Image Segmentation For Recognition and By Recognition (ICLR 2024 · similarity 0.81)
- Breaking the Reclustering Barrier in Centroid-based Deep Clustering (ICLR 2025 · 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.