Unsupervised EM-based joint optimization of interest point detector and descriptor via probability formulations of sparsity, repeatability and discriminability, yielding Property Network that outperforms SOTA on matching benchmarks without retraining.
Lift: Learned invariant feature transform
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2019 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Unsupervised Learning Framework of Interest Point Via Properties Optimization
Unsupervised EM-based joint optimization of interest point detector and descriptor via probability formulations of sparsity, repeatability and discriminability, yielding Property Network that outperforms SOTA on matching benchmarks without retraining.