Industrial Computerized Tomography Images Segmentation Based on Cellular Neural Networks
In this paper, we aimed at segmenting industrial computerized tomography images grounded in edge information. Most researchers focused on the edges of binary images using cellular neural network. We have expanded the scope to gray level images. Thus two groups of cellular neural network were designed to obtain closed edges. One was used to convert the gray level images to binary ones, the other to extract edges. In addition, we combined with contrastenhancement. Subsequently, tracing the obtained edges accomplished segmentation. Simulation experiments on a series of engine images demonstrate our methods can not only batch process threshold segmentation without choice threshold, but also extract the fine edges .A comparison to state-of-the-art methods show ours are easy to follow with good results.
C. J. Liu and X. L. Wu, "Industrial Computerized Tomography Images Segmentation Based on Cellular Neural Networks", Applied Mechanics and Materials, Vols. 66-68, pp. 2228-2235, 2011