Perceiving Systems Autonomous Motion Conference Paper 2016

Superpixel Convolutional Networks using Bilateral Inceptions

Gadde

In this paper we propose a CNN architecture for semantic image segmentation. We introduce a new “bilateral inception” module that can be inserted in existing CNN architectures and performs bilateral filtering, at multiple feature-scales, between superpixels in an image. The feature spaces for bilateral filtering and other parameters of the module are learned end-to-end using standard backpropagation techniques. The bilateral inception module addresses two issues that arise with general CNN segmentation architectures. First, this module propagates information between (super) pixels while respecting image edges, thus using the structured information of the problem for improved results. Second, the layer recovers a full resolution segmentation result from the lower resolution solution of a CNN. In the experiments, we modify several existing CNN architectures by inserting our inception modules between the last CNN (1 × 1 convolution) layers. Empirical results on three different datasets show reliable improvements not only in comparison to the baseline networks, but also in comparison to several dense-pixel prediction techniques such as CRFs, while being competitive in time.

Author(s): Raghudeep Gadde and Varun Jampani and Martin Kiefel and Daniel Kappler and Peter Gehler
Book Title: European Conference on Computer Vision (ECCV)
Year: 2016
Month: October
Series: Lecture Notes in Computer Science
Publisher: Springer
Project(s):
Bibtex Type: Conference Paper (inproceedings)
Event Name: 14th European Conference on Computer Vision
Event Place: Amsterdam, The Netherlands
Electronic Archiving: grant_archive
Attachments:

BibTex

@inproceedings{gadde16bilateralinception,
  title = {Superpixel Convolutional Networks using Bilateral Inceptions},
  booktitle = {European Conference on Computer Vision (ECCV)},
  abstract = {In this paper we propose a CNN architecture for semantic image segmentation. We introduce a new “bilateral inception” module that can be inserted in existing CNN architectures and performs bilateral filtering, at multiple feature-scales, between superpixels in an image. The feature spaces for bilateral filtering and other parameters of the module are learned end-to-end using standard backpropagation techniques. The bilateral inception module addresses two issues that arise with general CNN segmentation architectures. First, this module propagates information between (super) pixels while respecting image edges, thus using the structured information of the problem for improved results. Second, the layer recovers a full resolution segmentation result from the lower resolution solution of a CNN. In the experiments, we modify several existing CNN architectures by inserting our inception modules between the last CNN (1 × 1 convolution) layers. Empirical results on three different datasets show reliable improvements not only in comparison to the baseline networks, but also in comparison to several dense-pixel prediction techniques such as CRFs, while being competitive in time.},
  series = {Lecture Notes in Computer Science},
  publisher = {Springer},
  month = oct,
  year = {2016},
  slug = {gadde16bilateralinception},
  author = {Gadde, Raghudeep and Jampani, Varun and Kiefel, Martin and Kappler, Daniel and Gehler, Peter},
  month_numeric = {10}
}