Multi-resolution Image Segmentation via the Min-lifting Scheme and Watersheds

摘要

A multi-resolution image segmentation algorithm is proposed, which combines the min-lifting scheme with watershed transform. In order to overcome the over-segmentation of watershed transform in the original form, the algorithm uses the min-lifting scheme to pre-process the image and reduce the number of the minimums. Then the algorithm applies the watershed transform to the processed image and gets the segmentation result. The experimental results show the algorithm can decrease the number of regions and overcome the over-segmentation of watershed transform. The segmentation result is greatly improved using the multi-resolution image segmentation algorithm.