Automatic segmentation is an essential problem in biomedical imaging. It is still an open problem to automatically
segment biomedical images with complex structures and compositions. This paper proposes a novel algorithm called
Gradient-Intensity Clusters and Expanding Boundaries (GICEB). The algorithm attempts to solve the problem with
considerations of the image properties in intensity, gradient, and spatial coherence in the image space. The solution is
achieved through a combination of using a two-dimensional histogram, domain connectivity in the image space, and
segment region growing. The algorithm has been tested on some real images and the results have been evaluated.
Determining the neural connectivity of brain is an essential problem in neuroscience and the fluorescent imaging technique is a very useful to study this problem. In this technique, a real brain (typically of rat) is injected with a fluorescent dye and then sectioned into thin slices. Each slice is then exposed to illumination and a high-resolution image is captured. The areas in a slice that are impacted by the dye generate strong brightness due to fluorescence, and these regions reveal useful information on the neural connectivity. However, it is challenging to automatically register the image series. In this paper, we propose effective methods for the registration of fluorescent neural images. Our approach is based on the edge features of images. First, we use an effective method for edge detection. Then we adopt multi-level pattern recognition using clustering algorithms with the Mahalanobis distance criteria to isolate individual features. Finally, we adopt an elastic registration scheme using the thin-plate spline algorithm to solve the multivariate interpolation problem. Once all images are registered, we apply an elliptic weighted average (EWA) splatting technique for volume visualization. Our rendered results clearly display the 3D structures of the neural connectivity.