Adaptable physics-based super-resolution for electron backscatter diffraction maps

Abstract In computer vision, single-image super-resolution (SISR) has been extensively explored using convolutional neural networks (CNNs) on optical images, but images outside this domain, such as those from scientific experiments, are not well investigated. Experimental data is often gathered non-optical methods, which alters the metrics for image quality. One example electron backscatter diffraction (EBSD), a materials characterization technique that maps crystal arrangement in solid materials, provides insight into processing, structure, and property relationships. We present broadly adaptable approach applying state-of-art SISR to generate super-resolved EBSD orientation maps. This includes quaternion-based recognition, loss functions consider rotational effects crystallographic symmetry, an inference pipeline convert network output established visualization formats The ability physically accurate, high-resolution with enables high-throughput broadens capture capabilities three-dimensional experimental datasets.