Interactive shape co-segmentation via label propagation

• We provide an intuitive user interface. For the users to directly assign labels is more expressive than the pairwise constraints.
• We introduce an inductive extension of our algorithm to deal with out-of-sample data.
• We can achieve error-free results depending on the input dataset and the labels given by the users.
• Our approach is graph-based, but requires no extra eigen-decomposition, which is different from the unsupervised methods [7] and [8].

In this paper, we present an interactive approach for shape co-segmentation via label propagation. Our intuitive approach is able to produce error-free results and is very effective at handling out-of-sample data. Specifically, we start by over-segmenting a set of shapes into primitive patches. Then, we allow the users to assign labels to some patches and propagate the label information from these patches to the unlabeled ones. We iterate the last two steps until the error-free consistent segmentations are obtained. Additionally, we provide an inductive extension of our framework, which effectively addresses the out-of-sample data. The experimental results demonstrate the effectiveness of our approach.

The pipeline of the proposed method. Given a set of shapes, we first over-segment them to primitive patches, and based on which we construct the neighborhood graph. Then, the users are required to label some patches in the light of their intention. Finally, we propagate the label information of the labeled patches to the unlabeled ones, and obtain the co-segmentation results. The user can iterate the last two steps until satisfactory results have been achieved. We also provide an inductive extension of our approach to handle the out-of-sample data.Figure optionsDownload high-quality image (265 K)Download as PowerPoint slide