Skip to main content

Table 2 Validation set results for 11 classes at SNR = 0.6 in EMPIAR-10028

From: Simcryocluster: a semantic similarity clustering method of cryo-EM images by adopting contrastive learning

Metric

Top-1

Top-5

NMI

Pretext+K-means

0.57

–

0.63

SCAN

0.56

0.86

0.61

SimCryoCluster (Ours)

0.76

0.93

0.67

  1. The results with K-means were obtained using the pretext features from Simclr [30]. We provide the results obtained by our baseline(SCAN) [31],and the results we have improved on the model