Permutohedral Lattice CNNs

2015

Conference Paper

ei

ps


This paper presents a convolutional layer that is able to process sparse input features. As an example, for image recognition problems this allows an efficient filtering of signals that do not lie on a dense grid (like pixel position), but of more general features (such as color values). The presented algorithm makes use of the permutohedral lattice data structure. The permutohedral lattice was introduced to efficiently implement a bilateral filter, a commonly used image processing operation. Its use allows for a generalization of the convolution type found in current (spatial) convolutional network architectures.

Author(s): Martin Kiefel and Varun Jampani and Peter V. Gehler
Book Title: ICLR Workshop Track
Year: 2015
Month: May
Day: 7-9

Department(s): Empirical Inference, Perceiving Systems
Research Project(s): Learning to infer
Bibtex Type: Conference Paper (inproceedings)
Paper Type: Conference

Event Name: ICLR 2015
Event Place: San Diago
State: Published
URL: http://arxiv.org/abs/1412.6618

Links: pdf

BibTex

@inproceedings{kiefel_iclr_2015,
  title = {Permutohedral Lattice CNNs},
  author = {Kiefel, Martin and Jampani, Varun and Gehler, Peter V.},
  booktitle = {ICLR Workshop Track},
  month = may,
  year = {2015},
  url = {http://arxiv.org/abs/1412.6618},
  month_numeric = {5}
}