Medical image segmentation based on fast region connecting

An image segmentation approach is presented that merges watershed segmentation regions with the nearest neighbor connecting tree (NNCT). Firstly a dilation-erosion contrast enhancement processing is used as a preprocessing stage to obtain an accurate estimate of the image borders. Then the maker-controlled watershed transform is applied to produce an initial partitioning of the image into primitive regions. Lastly watershed regions are merged by constructing the NNCT to produce the last segmentation. In the latter process, the seed is introduced and the largest route connectedness is computed between the seed and every node in the route of the region adjacency graph (RAG). Simultaneously, a faster algorithm based on the prior principle of largest route connectedness is produced to create the NNCT, due to which processing steps are drastically reduced. The segmentation approach is applied to lung extraction in Computerized Tomography (CT) images. The results show the efficiency of the algorithm for medical image segmentation.