A controlled benchmark of 22 ImageNet-1K self-supervised models on 8 mid-level vision tasks shows a generally weak and task-dependent correlation with high-level performance.
Unsupervised learning of visual features by contrasting cluster assignments, 2021
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Probing the Mid-level Vision Capabilities of Self-Supervised Learning
A controlled benchmark of 22 ImageNet-1K self-supervised models on 8 mid-level vision tasks shows a generally weak and task-dependent correlation with high-level performance.