LSD-C: linearly separable deep clusters
We present LSD-C, a novel method to identify clusters in an unlabeled dataset. Our algorithm first establishes pairwise connections in the feature space between the samples of the minibatch based on a similarity metric. Then it regroups in clusters the connected samples and enforces a linear separat...
| Glavni autori: | , , , , |
|---|---|
| Format: | Conference item |
| Jezik: | English |
| Izdano: |
IEEE
2021
|