Geometry-Aware Unsupervised Domain Adaptation for Stereo Matching
Hiroki Sakuma,Yoshinori Konishi,Hiroki Sakuma,Yoshinori Konishi
Recently proposed DNN-based stereo matching methods that learn priors directly from data are known to suffer a drastic drop in accuracy in new environments. Although supervised approaches with ground truth disparity maps often work well, collecting them in each deployment environment is cumbersome and costly. For this reason, many unsupervised domain adaptation methods based on image-to-image tran...
