A Novel Methodology Based On Multi-Level Immersion Watershed

Segmentation is a fundamental problem in the image analysis. The general segmentation problem is the process of partitioning an image or data-set into a number of homogeneous segments. Although the methods of image segmentation have been improved significantly recently, it is still a very difficult problem in practice.The watershed transform is a popular segmentation method coming from the field of mathematical morphology. The intuitive description of this transform is quite simple: a 2D image can be considered as a topographic relief, where the height of each point is directly related to its gray level, and rain gradually falling on the terrain, and then the watersheds are the lines that separate the"lakes"that form.The watershed transform has been widely used in many fields of image processing. Due to the number of advantages that it possesses: it can be parallelized performance, and it produces a complete division of the image in separated regions even if the contrast is poor, thus avoiding the need for any kind of contours joining. However, some important drawbacks also exist, where the over-segmentation is the major one. Therefore, the improvement of the watershed algorithm to deal with the over-segmentation problem becomes an essential issue.To address the over-segmentation problem in the watershed processing, many studies have been made, which can be categorized into three parts: pre-processing, in-processing and post-processing. In this paper, a novel Multi-Level (ML) watershed algorithm is proposed to decrease the over- segmentation problem. The idea can be described as follows: in the...