Photons meeting: Bragg Diffraction Imaging – Overcoming Angular Uncertainty using Deep Learning

Huaiyu Chen from Synchrotron Radiation Research will give a talk titled: Bragg Diffraction Imaging: Overcoming Angular Uncertainty using Deep Learning

This is a rehearsal for his PhD defense on Sep 19^th. Abstract below!

After the talk, we will have a discussion and brainstorming about the Photons Friday meetings.

Abstract:

Advanced X-ray diffraction imaging techniques can uncover the internal structures of nanoscale crystalline materials—structures that are vital to technologies such as semiconductors and sensors. Using two powerful methods, Bragg Coherent Diffraction Imaging (BCDI) and scanning X-ray diffraction (nano-XRD), this work maps how atoms shift and strain deep within a material, all without cutting or damaging the sample. A major focus of the research is addressing one of BCDI’s key challenges: its sensitivity to angular motion and instability during experiments. To overcome this, I have developed a maximum likelihood correction method to mitigate angular distortions. Building on this, a deep learning approach was introduced to further enhance the method’s capability to fully unknown angles, potentially enabling BCDI to operate under much more complex scenarios.