top of page

We are committed to open source and to sharing the code developed within the scope of our research. 

A few examples are listed here and we refer to our GitHub repositories for more details

https://github.com/YijinLiu-Lab

TXM-Wizard is a toolbox for processing X-ray transmission image data collected using the Xradia TXM system.

​

References:
1) TXM-Wizard: a program for advanced data collection and evaluation in full-field transmission X-ray microscopy, Journal of Synchrotron Radiation, 19, 281-287. (2012)

2) 3D elemental sensitive imaging using transmission X-ray microscopy, Analytical and Bioanalytical Chemistry, Volume 404, Issue 5, pp 1297-1301 (2012)

3) Three-dimensional imaging of chemical phase transformations at the nanoscale with full-field transmission X-ray microscopy, Journal of Synchrotron Radiation, 18, 773-781. (2011)

TXM_Wizard_logo.gif

LIBNet is an implementation of Mask R-CNN on Python 3, Keras, and TensorFlow, adapted from matterport/Mask_RCNN , for the instance segmentation of NMC cathode particles in Lithium-ion battery electrode. The model generates bounding boxes and segmentation masks for each instance of an object in the image. It's based on Feature Pyramid Network (FPN) and a ResNet101 backbone.

Reference: 

Machine-learning-revealed statistics of the particle-carbon/binder detachment in lithium-ion battery cathodes, Nature Communications 11, 2310 (2020).

​

LIBNet.jpeg

This program is for conducting clustering of the EXAFS data from transmission X-ray microscopy (TXM)


   •    The post edge region (extended X-ray absorption fine structure, EXAFS), which can offer valuable information on the structure ordering/disordering and atomic bonding, is often not the focus in TXM spectro-microscopy studies. This is largely due to the limitation in the signal-to-noise ratio of the local spectra for EXAFS data reduction and fitting.
   •    Combining a machine-learning-based data classification method, we effectively reduce the dimensionality and complexity of the EXAFS data for detailed quantitative fitting.

​

Reference:

Understanding the mesoscale degradation in nickel-rich cathode materials through machine-learning-revealed strain-redox decoupling, ACS Energy Lett. 6, 687-693 (2021).

EXAFS_clustering.gif

Corresnet is a deep-learning-based image jitter correction method for synchrotron nano-resolution tomographic reconstruction.

Reference: 

Deep-learning-based image registration for nano-resolution tomographic reconstruction, Journal of Synchrotron Radiation, DOI:10.1107/S1600577521008481 (2021)

​

Corresnet.png

This program is for conducting clustering of the Nanodiffraction data from HXN beamline of NSLS-II

​

  • Lattice defects, e.g., dislocations and grain boundaries, critically impact the properties of crystalline battery cathode materials.

  • A long-standing challenge is to probe the meso-scale heterogeneity and evolution of lattice defects with sensitivity to atomic-scale details.

  • We tackle this issue with a unique combination of X-ray nanoprobe diffractive imaging and advanced machine learning techniques.

  • These results pave a direct way to the understanding of crystalline battery materials’ response under external stimuli with high fidelity, which provides valuable empirical guidance to defect-engineering strategies for improving the cathode materials against aggressive battery operation.

​

Reference:

Probing lattice defects in crystalline battery cathode using hard X-ray nanoprobe with data-driven modeling, Energy Storage Materials 45, 647-655 (2022).

NanoDiffraction.png
bottom of page