抄録
We consider the problem of approximating a function on a d-dimensional voxel grid by a unimodal function to minimize the L2 approximation error with respect to a given measure distribution on the grid. The output unimodal function gives a layered structure on the voxel grid, and we give efficient algorithms for computing the optimal approximation under a reasonable assumption on the shape of each horizontal layer. Our main technique is a dominating cut algorithm for a graph.
本文言語 | English |
---|---|
ページ(範囲) | 238-248 |
ページ数 | 11 |
ジャーナル | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
巻 | 3106 |
出版ステータス | Published - 2004 12 1 |
ASJC Scopus subject areas
- Theoretical Computer Science
- Computer Science(all)