Skip to main content
. 2018 May 31;41(7):1559–1572. doi: 10.1109/TPAMI.2018.2840695

Fig. 1.

Fig. 1.

Overview of the proposed interactive segmentation framework with two stages. Stage 1: P-Net automatically proposes an initial segmentation. Stage 2: R-Net refines the segmentation with user interactions indicating mis-segmentations. CRF-Net(f) is our proposed back-propagatable CRF that uses freeform pairwise potentials. It is extended to be CRF-Net(fu) that employs user interactions as hard constraints.