Multivariate watershed segmentation of compositional data
DGCI'09 Proceedings of the 15th IAPR international conference on Discrete geometry for computer imagery
Hi-index | 0.00 |
Abstract: In this paper, a new segmentation technique for multivalued images is elaborated. The technique makes use of the first fundamental form to access edge information of a multivalued image. On the obtained edge map, a watershed-based algorithm is applied. In order to remove noise or local texture, before segmentation, an anisotropic diffusion filter is applied, also making use of the first fundamental form. In this way, the entire procedure is applied using multivalued processing. Experiments are performed on colour images, medical multimodal images and multispectral satellite imagery. Segmentation results are compared to single-valued segmentation and filtering, applied to the intensity-only or the band-average images.