Feb 28 (Wed) @ 9:00am: “Enhancing Materials Microstructure Analysis with Physics-Informed Computer Vision,” Devendra Kumar Jangid, ECE PhD Defense

Date and Time
Harold Frank Hall (HFH), Room 4164 (ECE Conf. Room)


Collecting 3D microstructural information of materials poses significant challenges, being a process that is time-consuming and expensive. While advancements in serial sectioning instrumentation have expedited the acquisition of 3D data, the process of obtaining crystallographic information through electron backscatter diffraction (EBSD) imaging continues to be a bottleneck, limiting the overall rate of data collection. In this research, we explore physics-informed computer vision methods to generate high-resolution data. EBSD is a scanning electron microscope (SEM) imaging modality that maps crystal lattice orientation by analyzing diffraction patterns. EBSD maps are used to determine the microstructural properties such as texture, orientation gradients, phase distributions, and point-to-point orientation correlations, all of which are critical for understanding material performance. We developed a physics-inspired 3D deep learning framework to address the unique challenges associated with such EBSD maps, including rotational symmetry. The proposed quaternion convolution neural network (QCNN) with self-attention is used to super-resolve high resolution 3D microstructure data. We demonstrate, both qualitatively and quantitatively, that integrating the physics of microstructure into the deep learning architecture and loss function significantly reduces superresolution synthesis error compared to standard deep learning networks and loss functions.

Additionally, we propose a Generative Adversarial Network (GAN) framework known as M-GAN, which can be used to learn the morphologies of grains and synthesize realistic grains in microstructures. The creation of synthetic 3D grains represents a foundational step towards generating comprehensive synthetic 3D microstructures through deep learning techniques. The data and methods developed are available to the broader research community through the UCSB BisQue platform.


Devendra Kumar Jangid is a Ph.D. candidate in the Electrical and Computer Engineering department working with Professor Manjunath in the Vision Research Lab. His research focuses on the intersection of machine learning and material science, designing physics-based neural networks to generate high-resolution microstructure generation. He received his B.Tech in Electrical Engineering from Indian Institute of Technology, Jodhpur in 2015 and started his graduate studies at UC Santa Barbara in Fall 2018. His Ph.D. journey included internships in the Computational Camera Team at Samsung Research America.

Submitted by: Devendra Kumar Jangid <dkjangid@ucsb.edu>

Hosted by: Professor B.S. Manjunath