Abstract

In this paper, we propose a novel approach for ob ject co- segmentation in arbitrary videos by sampling, tracking and matching ob ject proposals via a Regulated Maximum Weight Clique (RMWC) extraction scheme. The proposed approach is able to achieve good seg- mentation results by pruning away noisy segments in each video through selection of ob ject proposal tracklets that are spatially salient and tem- porally consistent, and by iteratively extracting weighted groupings of ob jects with similar shape and appearance (with-in and across videos). The ob ject regions obtained from the video sets are used to initialize per- pixel segmentation to get the final co-segmentation results. Our approach is general in the sense that it can handle multiple ob jects, temporary oc- clusions, and ob jects going in and out of view. Additionally, it makes no prior assumption on the commonality of ob jects in the video collection. The proposed method is evaluated on publicly available multi-class video ob ject co-segmentation dataset and demonstrates improved performance compared to the state-of-the-art methods.