Rethinking Depthwise Separable Convolutions: How Intra-Kernel Correlations Lead to Improved MobileNets.

Daniel Haase Manuel Amthor

Published in: CoRR (2020)

Keyphrases

convolution kernel
fourier transform
improved algorithm
support vector
strongly correlated
mutual subspace method
multiscale
feature vectors
signal processing
kernel function
kernel methods
highly correlated
multiple kernel learning
correlation analysis
kernel regression
linearly separable