Segmentation by Local Binary Fitting Active Contour Model for Activated Carbon Fibers Material Microscopic Images
Jiang Zhao, Yu Michael Zhu, Jian Feng Yu · Advanced materials research · 2013
Many bubbles and pores are appeared on Activated Carbon Fibers (ACFs) material microscopic images. The morphology of ACFs surface image is complicated. Some widely used traditional methods are difficult to segment the object correctly. In this paper, an implicit active contour driven by local binary fitting energy is used to segment the objects for ACFs micro-images. This method is based on local image edge information to obtain optimal level set active contour model. Experimental results show that this active contour model is flexible for analyzing images with complex porous structure.