Dynamic Filter Networks release_5q5ptcahcnadpfkpyrizicfuji

by Bert De Brabandere, Xu Jia, Tinne Tuytelaars, Luc Van Gool

Released as a article .

2016  

Abstract

In a traditional convolutional layer, the learned filters stay fixed after training. In contrast, we introduce a new framework, the Dynamic Filter Network, where filters are generated dynamically conditioned on an input. We show that this architecture is a powerful one, with increased flexibility thanks to its adaptive nature, yet without an excessive increase in the number of model parameters. A wide variety of filtering operations can be learned this way, including local spatial transformations, but also others like selective (de)blurring or adaptive feature extraction. Moreover, multiple such layers can be combined, e.g. in a recurrent architecture. We demonstrate the effectiveness of the dynamic filter network on the tasks of video and stereo prediction, and reach state-of-the-art performance on the moving MNIST dataset with a much smaller model. By visualizing the learned filters, we illustrate that the network has picked up flow information by only looking at unlabelled training data. This suggests that the network can be used to pretrain networks for various supervised tasks in an unsupervised way, like optical flow and depth estimation.
In text/plain format

Archived Files and Locations

application/pdf  1.5 MB
file_liso7nbyvnebnbrbprrn6zqcvm
arxiv.org (repository)
web.archive.org (webarchive)
Read Archived PDF
Preserved and Accessible
Type  article
Stage   submitted
Date   2016-06-06
Version   v2
Language   en ?
arXiv  1605.09673v2
Work Entity
access all versions, variants, and formats of this works (eg, pre-prints)
Catalog Record
Revision: 05c05305-b6b2-40ba-8dd3-939473db6963
API URL: JSON