Fig. 1

The flowchart of the proposed model. The framework of our model consists of two stages: the training stage, which includes a depth CNN trained for feature learning and a generative process for saliency, and the testing stage. From a pair of RGB and depth images, our model extracts deep features using a colour CNN and depth CNN, respectively, and performs saliency prediction using the DMNB model [14] within a Bayesian framework. In this work, we perform experiments based on the NLPR dataset in [19]