Date of Award


Document Type

Open Access Dissertation


Computer Science and Engineering

First Advisor

Song Wang

Second Advisor

Lili Ju


Accurate grain segmentation on 3D superalloy images is very important in materials science and engineering. From grain segmentation, we can derive the underlying superalloy grains' micro-structures, based on which many important physical, mechanical and chemical properties of the superalloy samples can be evaluated. However, grain segmentation is usually a very challenging problem since: 1) even a small 3D superalloy sample may contain hundreds of grains; 2) carbides and noises may degrade the imaging quality; and 3) the intensity within a grain may not be homogeneous. In addition, the same grain may present different appearances, i.e. intensities, under different microscope settings. In practice, a 3D superalloy image may contain multichannel information where each channel corresponds to a specific microscope setting. In this work, we develop a Multichannel Edge-Weighted Centroidal Voronoi Tessellation (MCEWCVT) algorithm to effectively and robustly segment the superalloy grains in 3D multichannel superalloy images. MCEWCVT performs segmentation by minimizing an energy function which encodes both the mul tichannel voxel-intensity similarity within each cluster in the intensity domain and the smoothness of segmentation in the 3D image domain. Based on MCEWCVT, we further develop a Constrained Multichannel Edge-Weighted Centroidal Voronoi Tessellation (CMEWCVT) algorithm which can take manual segmentation on a small number of selected 2D slices as constraints from the problem domain, and incorporate them into the energy minimization process to further improve the segmentation accuracy. We quantitatively evaluate the MCEWCVT and the CMEWCVT algorithms on an authentic Ni-based dataset and two synthesized datasets against ground-truth segmentation. The qualitative and quantitative comparisons among the MCEWCVT, the CMEWCVT and 18 existing image segmentation algorithms on the authentic dataset demonstrate the effectiveness and robustness of the MCEWCVT and the CMEWCVT algorithms. In addition, the experiments on two synthesized datasets indicate that the optimal algorithm parameters found in the testing on the authentic dataset can be used on other superalloy datasets which have similar size and number of grains.