Sequential fuzzy cluster extraction by a graph spectral method
Pattern Recognition Letters
Normalized Cuts and Image Segmentation
IEEE Transactions on Pattern Analysis and Machine Intelligence
IEEE Computer Graphics and Applications
A Factorization Approach to Grouping
ECCV '98 Proceedings of the 5th European Conference on Computer Vision-Volume I - Volume I
"GrabCut": interactive foreground extraction using iterated graph cuts
ACM SIGGRAPH 2004 Papers
An Iterative Optimization Approach for Unified Image Segmentation and Matting
ICCV '05 Proceedings of the Tenth IEEE International Conference on Computer Vision - Volume 2
A Closed Form Solution to Natural Image Matting
CVPR '06 Proceedings of the 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition - Volume 1
Hi-index | 0.01 |
Two techniques are devised for a natural image matting method using semi-supervised object extraction. One is a guiding scheme for placement of user strokes specifying object or background regions and the other is a scheme of adjustment of object colors for conforming to composited background colors. We draw strokes at inhomogeneous color regions disclosed with an unsupervised cluster extraction method from which the semi-supervised algorithm is derived. Objects are composited with a new background after their color adjustment using a color transfer method with eigencolor mapping. This image matting method is then extended to videos. Strokes are drawn only in the first frame from which memberships are propagated to successive frames to extract objects in every frame. Performance of the proposed method is examined with images and videos experimented with existing matting methods.