Abstract
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, these filters share one common drawback. That is, the filter construction considers only the information in the guidance image and remains fixed (i.e., static guidance). When the local structures in the guidance and target images are not consistent, these techniques may transfer incorrect contents to the target image. our method can selectively transfer salient structures that are consistent in both guidance and target images.
网络结构如下:
The network architecture of our learning-based joint filters. The proposed model con