With ever-accumulating information of the material data accompanied with the development of computer technology and imaging technology, machine learning is necessary in analyzing TEM data. Machine learning is one of the most optimal tool for analyzing these large chunks of data in an efficient and delicate way.

Using machine learning technology, we are studying the structure-property relationship of materials by (1) analyzing the morphology of materials, (2) analyzing the crystal structure of materials from high resolution TEM (HRTEM) images or diffraction pattern (DP) image and (3) developing denoising technology to improve TEM resolution.