我们通过神经网络提取的特征计算各图像的内部显著图，并作为计算交叉显著性

The inside of each of the image saliency map we calculated by the neural network extracted feature, as calculated cross significant prior knowledge, then using the depth residuals network multi-image significant excavation, to obtain a plurality of images of the cross saliency map, and finally with within each image integration saliency map, a significant synergistic FIG.

We calculated the internal significance of each image by the characteristics extracted by the neural network, and as a priori knowledge of calculating the cross-significance, and then used the deep residual network to carry out multi-image significance mining, obtaining the cross-image significance map of multiple images, and finally merged with the internal significant figures of the images, and obtained the synergistic significance.

We use the features extracted by neural network to calculate the internal saliency map of each image, and as a prior knowledge to calculate the cross saliency. Then we use the depth residual network to mine the multi image saliency, and get the cross saliency map of multiple images. Finally, we fuse with the internal saliency map of each image to obtain the collaborative saliency map<br>