ViTNF swaps the MLP head of a vision transformer for a kernel-based neural-field classifier and reports very high generalized-category-discovery accuracy, though on a redefined easier protocol.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
ViTNF: Leveraging Neural Fields to Boost Vision Transformers in Generalized Category Discovery
ViTNF swaps the MLP head of a vision transformer for a kernel-based neural-field classifier and reports very high generalized-category-discovery accuracy, though on a redefined easier protocol.