|About this Abstract
||2018 TMS Annual Meeting & Exhibition
||Algorithm Development in Materials Science and Engineering
||Segmentation for Large Datasets of X-ray Microscopes by Using a Deep Convolutional Neural Network
||Xiaogang Yang, Vincent De Andrade, Francesco De Carlo, Nikhilesh Chawla, C. Shashank Kaira, William Scullin, Doga Gursoy
|On-Site Speaker (Planned)
Segmentation of X-ray microscopy images is always a challenging job, due to complicated image pattern and noise. The traditional methods of image process are not robust enough for these cases. The manual work is always the most reliable and popular solution, but cannot be applied to large datasets. Here we introduce a deep neural network approach to learn the mapping between the original images and manually segmented images. The trained network can be used to make the automatic segmentation for the large datasets. We validate this algorithm with transmission X-ray microscopy (TXM) images of an Al-Cu alloy. The algorithm can emulate the way of human to segment the X-ray images with a reliable quality. It can speed up the segmentation process to be of the magnitude of 10^2 to 10^3 faster than the human force.
||Planned: Supplemental Proceedings volume