Scene understanding is to predict a class label at each pixel of an image. In this study, we propose a semantic segmentation\nframework based on classic generative adversarial nets (GAN) to train a fully convolutional semantic segmentation model along\nwith an adversarial network. To improve the consistency of the segmented image, the high-order potentials, instead of unary or\npairwise potentials, are adopted. We realize the high-order potentials by substituting adversarial network for CRF model, which\ncan continuously improve the consistency and details of the segmented semantic image until it cannot discriminate the segmented\nresult from the ground truth. A number of experiments are conducted on PASCAL VOC 2012 and Cityscapes datasets, and the\nquantitative and qualitative assessments have shown the effectiveness of our proposed approach.
Loading....