SCAN-clustering

2000년 도입 · 논문 3편에서 사용

SCAN automatically groups images into semantically meaningful clusters when ground-truth annotations are absent. SCAN is a two-step approach where feature learning and clustering are decoupled. First, a self-supervised task is employed to obtain semantically meaningful features. Second, the obtained features are used as a prior in a learnable clustering approach. Image source: Gansbeke et al.

출처: SCAN: Learning to Classify Images without Labels

소개 논문: SCAN: Learning to Classify Images without Labels