2016

Harmonic Networks: Deep Translation and Rotation Equivariance

Worrall, Daniel E., Garbin, Stephan J., Turmukhambetov, Daniyar et al.

Understand

Translating or rotating an input image should not affect the results of many computer vision tasks.

Convolutional neural networks (CNNs) are already translation equivariant: input image translations produce proportionate feature map translations.
This is not the case for rotations.
Global rotation equivariance is typically sought through data augmentation, but patch-wise equivariance is more difficult.

Built on

Handwritten digit recognition with a back-propagation network
Y. LeCun, B. E. Boser, J. S. Denker, D. Henderson, R. E. Howard, W. E. Hubbard, and L. D. Jackel · 1989
Earlier work this paper cites.
The design and use of steerable filters
W. T. Freeman and E. H. Adelson · 1991
Earlier work this paper cites.
Rotation-invariant neoperceptron
B. Fasel and D. Gatica-Perez · 2006
Earlier work this paper cites.
An empirical evaluation of deep architectures on problems with many factors of variation
H. Larochelle, D. Erhan, A. C. Courville, J. Bergstra, and Y. Bengio · 2007
Earlier work this paper cites.
Actionable information in vision
S. Soatto · 2009
Earlier work this paper cites.
Learning to represent spatial transformations with factored higher-order boltzmann machines
R. Memisevic and G. E. Hinton · 2010
Earlier work this paper cites.
Contour detection and hierarchical image segmentation
P. Arbelaez, M. Maire, C. Fowlkes, and J. Malik · 2011
Earlier work this paper cites.
Transforming auto-encoders
G. E. Hinton, A. Krizhevsky, and S. D. Wang · 2011
Earlier work this paper cites.
Imagenet classification with deep convolutional neural networks
A. Krizhevsky, I. Sutskever, and G. E. Hinton · 2012
Earlier work this paper cites.
2d/3d rotation-invariant detection using equivariant filters and kernel weighted mapping
K. Liu, Q. Wang, W. Driever, and O. Ronneberger · 2012
Earlier work this paper cites.
Learning rotation-aware features: From invariant priors to equivariant descriptors
U. Schmidt and S. Roth · 2012
Earlier work this paper cites.

Similar

Learning invariant representations with local transformations
K. Sohn and H. Lee · 2012
Cited alongside, same era.
Adam: A method for stochastic optimization
Original
D. P. Kingma and J. Ba · 2014
Cited alongside, same era.
Visual boundary prediction: A deep neural prediction network and quality dissection
J. J. Kivinen, C. K. I. Williams, and N. Heess · 2014
Cited alongside, same era.
Very deep convolutional networks for large-scale image recognition
Original
K. Simonyan and A. Zisserman · 2014
Cited alongside, same era.
Visualizing and understanding convolutional networks
M. D. Zeiler and R. Fergus · 2014
Cited alongside, same era.
Deepedge: A multi-scale bifurcated deep network for top-down contour detection
G. Bertasius, J. Shi, and L. Torresani · 2015
Cited alongside, same era.
Pixel-wise deep learning for contour detection
Original
J. Hwang and T. Liu · 2015
Cited alongside, same era.
Batch normalization: Accelerating deep network training by reducing internal covariate shift
S. Ioffe and C. Szegedy · 2015
Cited alongside, same era.
Pushing the boundaries of boundary detection using deep learning
Original
I. Kokkinos · 2015
Cited alongside, same era.
Deeply-supervised nets
C. Lee, S. Xie, P. W. Gallagher, Z. Zhang, and Z. Tu · 2015
Cited alongside, same era.
Understanding image representations by measuring their equivariance and equivalence
K. Lenc and A. Vedaldi · 2015
Cited alongside, same era.

Then

Deep roto-translation scattering for object classification
E. Oyallon and S. Mallat · 2015
Later among the works it cites.
Deepcontour: A deep convolutional feature learned by positive-sharing loss for contour detection
W. Shen, X. Wang, Y. Wang, X. Bai, and Z. Zhang · 2015
Later among the works it cites.
Holistically-nested edge detection
S. Xie and Z. Tu · 2015
Later among the works it cites.
Group equivariant convolutional networks
Original
T. S. Cohen and M. Welling · 2016
Closest in time.
Steerable cnns
Original
T. S. Cohen and M. Welling · 2016
Closest in time.
Exploiting cyclic symmetry in convolutional neural networks
Original
S. Dieleman, J. De Fauw, and K. Kavukcuoglu · 2016
Closest in time.
Structured receptive fields in cnns
Original
J. Jacobsen, J. C. van Gemert, Z. Lou, and A. W. M. Smeulders · 2016
Closest in time.
TI-POOLING: transformation-invariant pooling for feature learning in convolutional neural networks
Original
D. Laptev, N. Savinov, J. M. Buhmann, and M. Pollefeys · 2016
Closest in time.
Learning covariant feature detectors
Original
K. Lenc and A. Vedaldi · 2016
Closest in time.
Learning rotation invariant convolutional filters for texture classification
Original
D. Marcos, M. Volpi, and D. Tuia · 2016
Closest in time.
A mathematical motivation for complex-valued convolutional networks
M. Tygert, J. Bruna, S. Chintala, Y. LeCun, S. Piantino, and A. Szlam · 2016
Closest in time.

Beyond the bibliography

alphaXiv searches the wider corpus for related work and actual follow-ups.

Open on alphaXiv

alphaXiv is searching for related work…

Reading the bibliography…