Fast Cubical Persistent Homology on 2D and 3D Images via Union-Find, Pruning, and Lookup Tables
About
We present Flash Cubical, a highly efficient computation of cubical persistence on a V-filtration for 2D and 3D images over $\mathbb{F}_2$. The implementation is built around three core ideas. First, cubical complexes satisfy properties that allow for the computation of persistence of the highest dimension via union-find and duality. Second, pruning of certain edges allows for a fast and efficient implementation of union-find. Third, the use of a lookup table, which exploits the regularity of cubical complexes to pre-compute local information. This avoids the need to compute local information at run time. To the best of our knowledge, this is the most efficient implementation of cubical persistence with a V-filtration, both in terms of time and memory costs. Although the paper focuses on persistence for V-filtration cubical complexes, the underlying ideas generalise naturally to T-filtrations on cubical complexes and suggest promising directions for other complexes.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| V-filtration computation | Lena 256 x 256 | Execution Time (s)0.01 | 5 | |
| V-filtration computation | Synthetic 2D medium (316 x 316) | Execution Time (s)0.02 | 5 | |
| V-filtration computation | Synthetic 2D large 1024 x 1024 | Execution Time (s)0.11 | 5 | |
| V-filtration computation | DIV2K 1024 x 1024 | Execution Time (s)0.09 | 5 | |
| V-filtration computation | Synthetic 3D medium (46^3) | Execution Time (s)0.06 | 3 | |
| V-filtration computation | Fuel 64^3 | Execution Time (s)0.03 | 3 | |
| V-filtration computation | Synthetic 3D large (128^3) | Time (s)2.21 | 3 | |
| V-filtration computation | Bonsai 128^3 | Time (s)0.53 | 3 | |
| V-filtration computation | Aneurism 128^3 | Time (s)0.28 | 3 |