Image Processing and Analysis of 3-D Microscopy Data

摘要

While there are many microstruc-tural parameters that can be measured from a planar two-dimensional (2-D) section through a material, there are many measurements that require knowledge of the full three-dimensional (3-D) micro structure, such as true size and shape of individual objects, connectivity and interfacial curvatures. Serial sectioning and reconstruction can reveal the 3-D microstructure but are often considered to be time consuming and labor intensive. However, what is not often realized is that the majority of the time invested in serial sectioning is spent in the image segmentation, wherein individual objects are digitally identified. This article reviews the current state of image segmentation and novel analysis within 3-D materials science. We will also briefly discuss the future possibilities for more efficient segmentation of digital images for a broader range of materials.