Convolutional neural network for saliency detection in images
Hooman Misaghi, Reza Askari Moghadam, Kurosh Madani · 2018
In this paper a fully convolutional neural network is proposed for detecting visual saliency in images. In this approach no down sampling or up sampling is used thus no information is lost and salient region is detected at pixel level. This Convolutional Neural Network is able to detect multiple salient regions with any size. This method is applicable to any input image size. The proposed CNN structure and simulation results are presented.