Hyperpixel Flow matches images of the same object class by voting with selected multi-layer CNN features, setting new state-of-the-art scores and introducing the SPair-71k benchmark.
Neighbourhood con- sensus networks
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2019 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Hyperpixel Flow: Semantic Correspondence with Multi-layer Neural Features
Hyperpixel Flow matches images of the same object class by voting with selected multi-layer CNN features, setting new state-of-the-art scores and introducing the SPair-71k benchmark.