[21a-B01-1~10] 23.1 Joint Session N "Informatics"

[21a-B01-8] Noise Reduction of Electron Hologram Using Sparse Coding

〇(P)Satoshi Anada¹, Nomura Yuki^2,3, Tsukasa Hirayama¹, Kazuo Yamamoto¹ (1.JFCC, 2.Panasonic, 3.Nagoya Univ.)

Keywords:Machine Learning, Image Processing, Electron Microscopy

Electron holography is a powerful technique to visualize electric and magnetic fields of a specimen by analyzing an electron interference pattern (hologram). The precision of electron holography depends directly on the noise in the hologram. In the present study, we have demonstrated that sparse coding, an image processing technique, effectively reduces the noise in the holograms of a GaAs p-n junction specimen recorded with short exposure times.