Abstract | ||
---|---|---|
We propose a new model for grain defect detection based on the theory of lattice metric space [7]. The lattice metric space (L, d(L)) shows outstanding advantages in representing lattices. Utilizing this advantage, we propose a new algorithm, Lattice clustering algorithm (LCA). After over-segmentation using regularized k-means, the merging stage is built upon the lattice equivalence relation. Since LCA is built upon (L, d(L)), it is robust against missing particles, deficient hexagonal cells, and can handle non-hexagonal lattices without any modification. We present various numerical experiments to validate our method and investigate interesting properties. |
Year | DOI | Venue |
---|---|---|
2019 | 10.1007/978-3-030-22368-7_30 | Lecture Notes in Computer Science |
Field | DocType | Volume |
Combinatorics,Equivalence relation,Lattice (order),Hexagonal crystal system,Metric space,Cluster analysis,Merge (version control),Physics | Conference | 11603 |
ISSN | Citations | PageRank |
0302-9743 | 0 | 0.34 |
References | Authors | |
0 | 2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Yuchen He | 1 | 0 | 1.01 |
Sung Ha Kang | 2 | 430 | 29.39 |