Abstract | ||
---|---|---|
Previous image segmentation methods are mainly based on region or contours. The former category of models are computationally expensive, while the latter approaches require lots of user interactions. In this paper, we propose a novel interactive image segmentation method, which makes a combination of the two models. By adopting the advantage of respective models, our method can produce high quality segmentations with little user interaction and achieve a surprisingly high efficiency. Specifically, we first obtain a coarse segmentation on a reduced scale using the classical graph cut method. Then we refine the boundary region on finer scales using active contours based method iteratively. The experimental results show that our method can produce better segmentation results to state-of-the-art while greatly reducing user interactions and processing time. We believe the proposed method could greatly improve user experiences in real applications. |
Year | DOI | Venue |
---|---|---|
2015 | 10.1145/2808492.2808547 | ICIMCS |
Field | DocType | Citations |
Cut,Computer vision,Scale-space segmentation,Pattern recognition,Segmentation,Computer science,Scale space,Segmentation-based object categorization,Image segmentation,Region growing,Artificial intelligence | Conference | 0 |
PageRank | References | Authors |
0.34 | 9 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Ling Ge | 1 | 0 | 1.35 |
Ran Ju | 2 | 108 | 7.87 |
Gangshan Wu | 3 | 275 | 36.63 |