Learning on the edge: investigating boundary filters in CNNs

DOI10.1007/S11263-019-01223-YMaRDI QIDQ2193548zbMATH OpenOpenAlexWikidataFDO

Authors Carlo Innamorati, Tobias K. S. Ritschel, Tim Weyrich, Niloy J. Mitra

Publication date 20 August 2020

Published in International Journal of Computer Vision (Search for Journal in Brave)

Copyright license Creative Commons Attribution 4.0 International

Full work available at URL https://arxiv.org/abs/1805.03106

zbMATH Keywords

boundary conditionsboundary rulesconvolutional neural networksdeep learning

Mathematics Subject Classification ID

Artificial neural networks and deep learning (68T07) Machine vision and scene understanding (68T45)

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).

Recommendations

• The influence of the boundary conditions on the robustness of a CNN
• Synthetic boundary conditions for image deblurring
• Networks for nonlinear diffusion problems in imaging
• Blind image inpainting with sparse directional filter dictionaries for lightweight CNNs
• A boundary condition based deconvolution framework for image deblurring

Cites work

• Deep learning

Cited in

(1)

• A deep learning modeling framework to capture mixing patterns in reactive-transport systems

Describes a project that uses

Uses Software

• ImageNet
• U-Net
• AlexNet
• Keras
• FlowNet

This page was built for publication: Learning on the edge: investigating boundary filters in CNNs

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2193548)