Learning on the edge: Explicit boundary handling in CNNs
Innamorati C, Ritschel T, Mitra NJ, Weyrich T (2019)
Publication Type: Conference contribution
Publication year: 2019
Publisher: BMVA Press
Conference Proceedings Title: British Machine Vision Conference 2018, BMVC 2018
Event location: Newcastle
Abstract
Convolutional neural networks (CNNs) handle the case where filters extend beyond the image boundary using several heuristics, such as zero, repeat or mean padding. These schemes are applied in an ad-hoc fashion and, being weakly related to the image content and oblivious of the target task, result in low output quality at the boundary. In this paper, we propose a simple and effective improvement that learns the boundary handling itself. At training-time, the network is provided with a separate set of explicit boundary filters. At testing-time, we use these filters which have learned to extrapolate features at the boundary in an optimal way for the specific task. Our extensive evaluation, over a wide range of architectural changes (variations of layers, feature channels, or both), shows how the explicit filters result in improved boundary handling. Consequently, we demonstrate an improvement of 5 % to 20 % across the board of typical CNN applications (colorization, de-Bayering, optical flow, and disparity estimation).
Authors with CRIS profile
Involved external institutions
How to cite
APA:
Innamorati, C., Ritschel, T., Mitra, N.J., & Weyrich, T. (2019). Learning on the edge: Explicit boundary handling in CNNs. In British Machine Vision Conference 2018, BMVC 2018. Newcastle, GB: BMVA Press.
MLA:
Innamorati, Carlo, et al. "Learning on the edge: Explicit boundary handling in CNNs." Proceedings of the 29th British Machine Vision Conference, BMVC 2018, Newcastle BMVA Press, 2019.
BibTeX: Download