[3Rin2] Interactive Session 1

[3Rin2-22] Noise reduction of live image in scanning electron microscope

〇Fuminori Uematsu¹, Masahiko Takei¹, Mitsuyoshi Yoshida¹ (1. JEOL Ltd.)

Keywords:generative adversarial networks, scanning electron microscope, noise reduction

A real time display called a live image is used to search for a region to be observed, when observing a sample with a scanning electron microscope (SEM). This live image is usually an image with high noise and poor visibility. This makes it difficult to find an appropriate observation area. Therefore, in this research, a noise reduction model by deep neural network was created with the aim of improving the visibility of live images. We incorporated this model into the SEM and succeeded in obtaining a live image with less noise. This makes it possible to efficiently search for the observation region.