Joint image filters can leverage the guidance image as a prior and transfer the structural details from the guidance image to the target image for suppressing noise or enhancing spatial resolution. Existing methods rely on various kinds of explicit filter construction or handdesigned objective functions. It is thus difficult to understand, improve, and accelerate them in a coherent framework. In this paper, we propose a learning-based approach to construct a joint filter based on Convolutional Neural Networks. In contrast to existing methods that consider only the guidance image, our method can selectively transfer salient structures that are consistent in both guidance and target images. We show that the model trained on a certain type of data, e.g., RGB and depth images, generalizes well for other modalities, e.g., Flash/Non-Flash and RGB/NIR images. We validate the effectiveness of the proposed joint filter through extensive comparisons with state-of-the-art methods.
|Title of host publication||Computer Vision - 14th European Conference, ECCV 2016, Proceedings|
|Editors||Bastian Leibe, Jiri Matas, Nicu Sebe, Max Welling|
|Number of pages||16|
|Publication status||Published - 2016|
|Event||14th European Conference on Computer Vision, ECCV 2016 - Amsterdam, Netherlands|
Duration: 2016 Oct 8 → 2016 Oct 16
|Name||Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)|
|Conference||14th European Conference on Computer Vision, ECCV 2016|
|Period||16/10/8 → 16/10/16|
Bibliographical noteFunding Information:
This work is supported in part by the NSF CAREER Grant #1149783, gifts from Adobe and Nvidia, and Office of Naval Research under grant N00014-16-1-2314.
© Springer International Publishing AG 2016.
All Science Journal Classification (ASJC) codes
- Theoretical Computer Science
- Computer Science(all)