Vapor depression segmentation and absorptivity prediction from synchrotron x-ray images using deep neural networks
Runbo Jiang
Carnegie Mellon University, USA
: Res J Opt Photonics
Abstract
The quantification of the amount of absorbed light is essential for understanding laser-material interactions and melt pool dynamics in order to minimize defects in additively manufactured metal components. The geometry of a vapor depression, also known as a keyhole, in melt pools formed during laser melting is closely related to laser absorptivity. This relationship was observed by the state-of-the-art in situ high speed synchrotron x-ray visualization and integrating sphere radiometry. These two techniques create a temporally resolved dataset consisting of keyhole images and the corresponding laser absorptivity. In this work, we developed two pipelines for energy absorption prediction. The one-stage approach adopts convolutional neural networks (ConvNets) to learn feature kernels automatically directly yields an absorption value for each x-ray image using the fully connected layer and regression layer. The two-stage approach first generates semantic image segmentation using ConvNets for keyholes, then extract the geometric keyhole features, and finally applies regression models. Motivation for the second approach is that many artificial intelligence tasks can be solved by carefully designing the right set of features to extract for the task and then feeding these features to a simple machine learning algorithm. This is especially applicable in this case, where we already have a clear understanding of which features are relevant and should be extracted. On the other hand, the main advantage of one-stage approach to the stage-stage approach is that it automatically detects the important features and make a prediction without any human supervision. We compare the two different absorption prediction pipelines in terms of accuracy, generalizability on unseen materials, and model interpretability. This work provides a new pathway for process-control and a real-time monitoring system of laserbased additive manufacturing.
Biography
Runbo Jiang is a final year Ph.D. student in the Department of Materials Science and Engineering at Carnegie Mellon University. She works at the intersection of materials science and machine learning. She has her expertise in applying machine learning models in solving complex materials problems. Her deep learning models on the prediction of laser absorption for synchrotron x-ray images create a new pathway for process-control and a real-time monitoring system of laserbased additive manufacturing. She is expected to graduate in Aug 2023 and is currently actively seeking Machine Learning Engineer (MLE) and Data Scientist (DS) positions. She is highly skilled in materials science, statistics, machine learning and software engineering. She is excited to bring her expertise and passion to companies where she can make a meaningful impact.
