A Model of Saliency-Based Visual Attention for Rapid Scene Analysis
IEEE Transactions on Pattern Analysis and Machine Intelligence
"GrabCut": interactive foreground extraction using iterated graph cuts
ACM SIGGRAPH 2004 Papers
Active segmentation for robotics
IROS'09 Proceedings of the 2009 IEEE/RSJ international conference on Intelligent robots and systems
Salient region detection and segmentation
ICVS'08 Proceedings of the 6th international conference on Computer vision systems
Saliency detection for content-aware image resizing
ICIP'09 Proceedings of the 16th IEEE international conference on Image processing
| Hi-index | 0.00 |
Thanks to many state-of-the-art researches on visual attention techniques, we have ability to locate or focus on salient objects automatically, which are usually the main characters of the scene. To robustly extract the salient object, this paper suggests a combination of visual attention techniques and active segmentation. Active segmentation has introduced an innovative idea about inside-out segmentation but we have to manually specify the initial fixations to make the algorithm work. So, visual attention techniques are used to grant active segmentation the automatic ability. To that purpose, this work also introduces a local voting method to identify the ideal fixation points and merging rules for an optimal presentation of output. The detection and segmentation results show that the proposed method can automatically detect and segment salient objects robustly while it also can totally avoid the appearance of scatter segments, which frequently occur due to segmentation with binarized saliency map. The performance of our proposed combination model on public salient object database outperforms the other tested methods in terms of precision, recall and F-measure.